Khaliq_Jibran_PhD_final_151014

Transcription

Khaliq_Jibran_PhD_final_151014
Effect of Doping and Defect Structures on Thermo Physical Properties of
Thermoelectric Materials
Khaliq, Jibran
The copyright of this thesis rests with the author and no quotation from it or information
derived from it may be published without the prior written consent of the author
For additional information about this publication click this link.
http://qmro.qmul.ac.uk/jspui/handle/123456789/8444
Information about this research object was correct at the time of download; we occasionally
make corrections to records, please therefore check the published record when citing. For
more information contact [email protected]
Effect of Doping and Defect
Structures on Thermo Physical
Properties of Thermoelectric
Materials
Jibran Khaliq
Submitted in partial fulfillment of the requirements
of the Degree of Doctor of Philosophy
School of Engineering and Materials Science,
Queen Mary, University of London
London, United Kingdom
October 2014
Declaration
I hereby declare that the present work is prepared solely by myself during the course
of my doctoral studies at the Queen Mary, University of London. It has not been
submitted anywhere for any award. Work of other people is fully acknowledged
according to standard referencing.
This thesis fully complies with the regulations set by the University of London and
the Queen Mary, University of London.
Jibran Khaliq
October 2014
Abstract
Abstract
Development of thermoelectric materials to date has focused on materials
that can operate at lower temperatures. However; there is now an increased need to
develop materials for higher temperature applications. In this research, medium to
high temperature oxide and non-oxide thermoelectric materials were fabricated and
characterized. For oxide thermoelectric materials, La4Ti4O14 and Sr4Nb4O14 were
chosen. These compounds are members of the homologous A4B4O14 series and
possess perovskite-like layered structure (PLS). PLS compounds have low thermal
conductivity due to a layered structure compared to the perovskite materials (e.g.
SrTiO3). These atomic scale layers help to reduce the thermal conductivity of PLS
compounds. Doping in PLS materials also creates atomic scale disorders. The effect
of acceptor-donor doping and oxidation-reduction on the thermal conductivity of
PLS ceramics were investigated in relation to mass contrast and compositional nonstoichiometry. High resolution TEM and XPS revealed that acceptor doping of
La4Ti4O14 produced nanoscale intergrowth regions of n=5 layered phase inside n=4
layered phase, while donor doping produced nanoscale intergrowth regions of n=3
layered structure. As a result of these nanoscale intergrowths, the thermal
conductivity value reduced by ~ 20% compared to the theoretical value. Pure
La4Ti4O14 has a thermal conductivity value of ~ 1.1 W/m.K which dropped to a
value of ~ 0.98 W/m.K in Sr doped La4Ti4O14 and ~ 0.93 W/m.K in Ta doped
La4Ti4O14. Pure Sr4Nb4O14 has a thermal conductivity value of ~ 1.05 W/m.K which
dropped to ~ 0.6 W/m.K after La doping. The factors influencing the thermal
conductivity of PLS compounds were also discussed.
Page i
Abstract
For non-oxide ceramics, CoSb3 was chosen due to its cage-like structure and
ideal for the application of Phonon Glass Electron Crystal Concept. The cage like
structure gives room to engineer its electrical and thermal properties without
affecting the other. For the first time, CoSb3 stuffed with Yb and substituted with Te
(YbyCoSb3-xTex) was synthesized by mechanical alloying and spark plasma
sintering. The electrical and thermal properties were characterized for pure and
doped material. A Seebeck coefficient value of ~ 160 µV/K was obtained at ~ 600800 K for Yb0.075CoSb2.85Te0.15. The electric resistivity dropped from ~ 1000 µΩm
for pure CoSb3 to ~ 9 µΩm for Yb0.075CoSb2.85Te0.15. Lattice thermal conductivity
was significantly reduced to a very low value of 1.17 W/m.K by the addition of Yb
atoms into CoSb2.85Te0.15 without significantly affecting its Seebeck coefficient and
electrical resistivity. This value is comparable to those produced by the costly
processing of nanostructured materials. A zT value of ~ 0.70 was obtained at 600 K.
This research has shown that by engineering the defect chemistry of
thermoelectric materials, it is possible to significantly reduce their thermal
conductivity without compromising their electrical properties.
Page ii
Acknowledgement
Acknowledgement
I would like to express my deepest gratitude to my Supervisor, Prof. Michael
John Reece for his excellent guidance, caring, patience, and providing me with an
excellent atmosphere for doing research which kept me motivated throughout my
PhD. He persuasively conveyed an interest in my work, and I am grateful for my
inclusion in his research group. I would also like to thank my secondary supervisor
Dr. Haixue Yan, whose critical analysis, illuminating discussions and friendly jokes
helped me to produce this piece of research.
I would like to thank Dr. Zofia Luklinska and Dr. Rory Wilson, for their
generous help. I would also like to thank Dr. Na Ni and Samuel Jackson for their
help with sample characterization.
I acknowledge various stimulating discussions, ideas and support provided
by member of our research group Dr. Giuseppe Viola, Dr. Salvatore Grasso, Dr.
Huanpo Ning, Dr. Zhipeng Gao, Chunchun Li, Harshit Porwal, Chen Chen, Kan
Chen and Qinghui Jiang without their synergistic excellence and ideas, I could
possibly not have reached this far.
A special thanks to my family. Words cannot express how grateful I am to
my mother and father for all of the sacrifices that they’ve made on my behalf. Their
prayers for me were what sustained me thus far. I would also like to thank my
brothers and sister for their endless love and support all through my career. I would
also like to thank my parents in law for their love, support and encouragement. I
would also like to extend my acknowledgement to all my friends who supported me
in writing, and encouraged me to strive towards my goal.
Page iii
Acknowledgement
At the end I would like to express appreciation to my beloved wife Amina
who spent sleepless nights with me and was always my support in the moments
when there was no one to answer my queries. Without any doubt, I dedicate this
thesis to my family.
Thanks for the financial support from QMUL and Nanoforce Technology
Limited, U.K.
Page iv
Tables of Contents
Table of Contents
Abstract ......................................................................................................................... i
Acknowledgement ..................................................................................................... iii
Table of Contents ......................................................................................................... v
Chapter I. Introduction ................................................................................................. 1
References ................................................................................................................ 4
Chapter II. Literature Review ...................................................................................... 6
2.1
Introduction to thermoelectrics...................................................................... 6
2.2
Efficiency of thermoelectric materials .......................................................... 7
2.3
Thermoelectric Materials............................................................................... 9
2.4
Oxide Thermoelectrics ................................................................................ 15
2.5
Perovskite Related Structures ...................................................................... 23
2.6
Thermal Conductivity.................................................................................. 35
2.7
Oxide Thermoelectric Modules ................................................................... 38
References .............................................................................................................. 40
Chapter III. Experimental Details .............................................................................. 50
3.1
Powder Preparation ..................................................................................... 50
3.2
Sintering by spark plasma sintering ........................................................... 52
3.3
Post Sintering treatment .............................................................................. 53
3.4
Characterization........................................................................................... 54
References .............................................................................................................. 60
Page v
Tables of Contents
Chapter IV. Characterization of La4Ti4O14±δ ............................................................. 61
4.1
Introduction ................................................................................................. 61
4.2
Experimental Details ................................................................................... 62
4.3
La4-xSrxTi4O14-
4.4
La4Ti4-xTaxO14+
±δ ........................................................................................ 79
4.5
La4Ti4-xNbxO14+
±δ ....................................................................................... 88
4.6
Conclusion ................................................................................................... 93
±δ ......................................................................................... 65
References .............................................................................................................. 94
Chapter V. Characterization of Sr4Nb4O14±δ .............................................................. 96
5.1
Introduction ................................................................................................. 96
5.2
Experimental Details ................................................................................... 97
5.3
Results and Discussions .............................................................................. 98
5.4
Conclusion ................................................................................................. 110
References ............................................................................................................ 111
Chapter VI. Thermal Conductivity of PLS Compounds .......................................... 112
6.1
Introduction ............................................................................................... 112
6.2
Experimental Details ................................................................................. 113
6.3
La4Ti4O14 ................................................................................................... 114
6.4
Sr4-xLaxNbxO14+
6.5
Discussions ................................................................................................ 141
6.6
Conclusion ................................................................................................. 148
±δ ..................................................................................... 135
Page vi
Tables of Contents
References ............................................................................................................ 149
Chapter VII. Thermoelectric Properties of Co-doped CoSb3................................... 150
7.1
Introduction ............................................................................................... 150
7.2
Experimental Details ................................................................................. 151
7.3
Results and Discussions ............................................................................ 152
7.4
Conclusion ................................................................................................. 161
References ............................................................................................................ 162
Chapter VIII. Conclusions and Future Work ........................................................... 165
8.1
Conclusion ................................................................................................. 165
8.2
Future Work............................................................................................... 168
List of my publications ............................................................................................ 170
Page vii
List of Figures and Tables
List of Figures
Figure 2.1: Typical thermoelectric modules: (a) for cooling; (b) for power
generation
Figure 2.2: (a) zT values of different thermoelectric materials; (b) development in
thermoelectrics over the past 20 years
Figure 2.3: (a) crystal structure of typical skutterudite Rh4Sb12 filled with In (red
balls), while blue balls represent Rh atoms and yellow balls Sb atoms; (b) shows a
cage formed by 12 Sb atoms with In filler inside
Figure 2.4: Effect of hall carrier concentration on electrical resistivity and Seebeck
coefficient
Figure 2.5: Variation of: (a) electrical conductivity; (b) Seebeck coefficient; (c)
thermal conductivity; (d) zT for Yb0.2Co4Sb12with temperature
Figure 2.6: Crystal structures of some of the layered cobalt oxides
Figure 2.7: Crystal Structure of ZnO (wurtzite)
Figure 2.8: Typical structure of In2O3 crystal
Figure 2.9: Influence of Sn and Ti in In2O3 on thermal conductivity and zT
Figure 2.10: Cubic perovskite unit cell. Red spheres represent the A cation, Blue
spheres represent the B cation and Grey spheres represent X anions
Figure 2.11: Perovskite crystal structure showing the oxygen
Figure 2.12: Typical structure of distorted CaMnO3
Figure 2.13: Schematic drawing of structure of Aurivillius compounds Bi2WO6
(n=1), SrBi2Ta2O9 (n=2) and Bi4Ti3O12 (n=3)
Figure 2.14: Crystal Structure of Dian Jacobson phase CsBiNb2O7
Page viii
List of Figures and Tables
Figure 2.15: Crystal structure of Sr3Ti2O7
Figure 2.16: Structural diagram of non-distorted AnBnO3n+2 projected along a axis
Figure 2.17: Thermal conductivity as a function of temperature for Bi4Ti3O12
Figure 2.18: Temperature dependence of power factor, thermal conductivity and
figure of merit for rare earth doped Sr3Ti2O7
Figure 2.19: Temperature dependence of thermal conductivity of various oxides
Figure 2.20: Crystal Structure of Sr4Nb4O14
Figure 2.21: Variation in: (a) thermal conductivity; (b) Seebeck coefficient; (c)
electrical resistivity with temperature for Sr3.6La0.4Nb4O14
Figure 2.22: Variation of room temperature thermal conductivity with increasing
amount of La
Figure 2.23: Thermal conductivity as a function of temperature for the textured and
randomly oriented polycrystalline Sr3.9La0.1Nb4O14
Figure 2.24: Crystal structure of pyrochlore La2Ti2O7
Figure 2.25: Temperature dependence of thermal conductivity in CoSb3-xTex
Figure 3.1: (a) QM-3SP4 Planetary ball mill machine (b) carbolite HTF 1800
Chamber furnace
Figure 3.2: Saffron scientific glove box used for powder handling under inert
environment
Figure 3.3: Spark Plasma Sintering furnace (FCT, Germany) with a schematic of its
chamber
Figure 3.4: Schematic representation of: (a) SPS Die set; (b) photo of the graphite
die during SPS process
Figure 3.5: Lab build (in china) electrical resistivity and Seebeck coefficient
measurement system
Page ix
List of Figures and Tables
Figure 3.6: Netzsch laser flash LFA 457
Figure 4.1: Typical processing parameters during sintering of: (a) La4Ti4O14±δ;
(b) La3.2Sr0.8Ti4O13.6±δ
Figure 4.2: Photos of ceramic disc: (a) after sintering; (b) after air annealing;
(c) after reduction
Figure 4.3: XRD patterns of the La4-xSrxTi4O14-
±δ:
(a) powders; (b) sintered; (c) air
annealed; (d) reduced samples
Figure 4.4: Typical SEM images of La3.2Sr0.8Ti4O13.6±δ powder after calcination:
(a) before ball milling; (b) after ball milling
Figure 4.5: SEM images of LST ceramics after polishing and etching:
(a)
La3.8Sr0.2Ti4O13.6±δ;
(b)
La3.6Sr0.4Ti4O13.6±δ;
(c)
La3.6Sr0.6Ti4O13.6±δ;
(d)
La3.2Sr0.8Ti4O13.6±δ
Figure 4.6: TEM images of air annealed: (a) La4Ti4O14±δ; (b) high magnification
image of La4Ti4O14±δ;(c) diffraction pattern of the air annealed La4Ti4O14±δ;
(d)
La3.2Sr0.8Ti4O13.6±δ
showing
the
large
defect
(intergrowth)
number;
(e) La3.2Sr0.8Ti4O13.6±δ; (f) La3.2Sr0.8Ti4O13.6±δ showing termination of intergrowths;
(g-h) La3.2Sr0.8Ti4O13.6±δ showing two regions of intergrowths
Figure 4.7: TEM images of reduced: (a) La4Ti4O14±δ; (b) high magnification image
of La4Ti4O14±δ; (c-d) La3.2Sr0.8Ti4O13.6±δ showing the large defect (intergrowth)
density;(e) a high magnification image of a typical region in La3.2Sr0.8Ti4O13.6±δ
showing intergrowths; (f) corresponding diffraction pattern of La3.2Sr0.8Ti4O13.6±δ
Figure 4.8: XPS spectrum of La4-xSrxTi4O14-
±δ:
(a) La 3d; (b) Ti 2p; (c) O1s; (d) Sr
3d
Figure 4.9: XRD patterns of the La4Ti4-xTaxO14±x: (a) powders; (b) sintered; (c) air
annealed; (d) reduced samples
Page x
List of Figures and Tables
Figure 4.10: TEM images of: (a) air annealed La4Ti3.6Ta0.4O14.2±δ; (b) high
magnification
image
of
air
annealed
La4Ti3.6Ta0.4O14.2±δ;
(c)
reduced
La4Ti3.6Ta0.4O14.2±δ; (d) high magnification image of reduced d La4Ti3.6Ta0.4O14±δ
(e) high magnification image of reduced d La4Ti3.6Ta0.4O14.2±δ with d spacing;
(f) diffraction pattern of high magnification image of reduced d La4Ti3.6Ta0.4O14±δ
Figure 4.11: XPS spectrum of La4Ti4-xTaxO14+
±δ:
(a) La 3d; (b) Ti 2p; (c) O1s; (d)
Ta 4f
Figure 4.12: XRD patterns of the La4Ti4-xNbxO14+
±δ:
(a) powders; (b) sintered; (c)
air annealed; (d) reduced samples
Figure 4.13: TEM images of: (a) air annealed La4Ti3.4Nb0.6O14.3±δ; (b) air annealed
La4Ti3.4Nb0.6O14.3±δ showing region of LaNbO4; (c) typical grain of reduced
La4Ti3.4Nb0.6O14.3±δ ceramic showing LaNbO4 grain; (d) corresponding diffraction
pattern of LaNbO4 grain; (e) a typical region showing cluster of La4Ti4O14 grains in
reduced ceramic; (f) a typical LaNbO4 grain showing planar defects in reduced
ceramic
Figure 5.1: XRD patterns of Sr4-xLaxNb4O14+
±δ:
(a) powder; (b) sintered; (c) air
annealed; (d) reduced samples
Figure 5.2: SEM images of Sr4-xLaxNb4O14±δ ceramics after polishing and etching
(a) Sr4Nb4O14; (b) Sr3.8La0.2Nb4O14.1±δ and; (c) Sr3.2La0.8Nb4O14.4±δ
Figure 5.3: TEM images of: (a) air annealed Sr4Nb4O14; (b) high magnification
image of air annealed Sr4Nb4O14±δ; (c) corresponding diffraction pattern for
Sr4Nb4O14±δ; (d) air annealed Sr3.2La0.8Nb4O14.4±δ; (e) air annealed Sr3.2La0.8Nb4O14.4±δ
showing regions of Sr4Nb4O14 and LaNbO4; (f) high magnification image of a single
LaNbO4 grain
Page xi
List of Figures and Tables
Figure 5.4: TEM images of: (a) reduced Sr4Nb4O14±δ; (b) high magnification image
of reduced Sr4Nb4O14±δ; (c-d) typical regions of reduced Sr3.2La0.8Nb4O14.4±δ; (e) high
magnification image of a typical Sr/Nb rich region
Figure 5.5: XPS spectrum of Sr3.6La0.4Nb4O14.2±δ: (a) Sr 3d; (b) Nb 3d; (c) O1s and;
(d) La3d
Figure 6.1: Variation in thermal diffusivity of La4-xSrxTi4O14±δ ceramics: (a) air
annealed; (b) reduced
Figure 6.2: Variation in specific heat capacity of sapphire with temperature
Figure
6.3:
Variation
in
specific
heat
capacity
of:
(a)
La4Ti4O14±δ;
(b) La3.2Sr0.8Ti4O13.6±δ ceramics with temperature
Figure 6.4: Variation in specific heat capacity of reduced La4Ti4O14±δ with
temperature
Figure 6.5: Variation in thermal conductivity of: (a) air annealed La4-xSrxTi4O13.6±δ;
(b) reduced La4-xSrxTi4O14-
±δ;
(c) La3.2Sr0.8Ti4O13.6±δ; (d) La4Ti4O14±δ with
temperature
Figure 6.6: Variation of thermal conductivityof reduced La4-xSrxTi4O14-
±δ
with Sr
content at 573 K
Figure 6.7: Variation in thermal diffusivity of La4Ti4-xTaxO14+
±δ
ceramics: (a) air
annealed; (b) reduced
Figure
6.8:
Variation
in
specific
heat
capacity
of:
(a)
La4Ti4O14±δ;
(b) La4Ti3.4Ta0.6O14.3±δ ceramics with temperature
Figure 6.9: Variation in thermal conductivity of: (a) air annealed La4Ti4-xTaxO14+
(b) reduced La4Ti4-xTaxO14+
±δ;
±δ;
(c) La4Ti4O14±δ; (d) La4Ti3.4Ta0.6O14.3±δ with
temperature
Page xii
List of Figures and Tables
Figure 6.10: Variation of thermal conductivity with Ta content for reduced
La4Ti4-xTaxO14+
±δ
at 573 K
Figure 6.11: Variation in thermal diffusivity of La4Ti4-xNbxO14+
±δ
ceramics: (a) air
annealed; (b) reduced
Figure
6.12:
Variation
in
specific
heat
capacity
of:
(a)
La4Ti4O14±δ;
(b) La4Ti3.4Nb0.6O14.3±δ ceramics with temperature
Figure
6.13:
La4Ti4-xNbxO14+
Variation
±δ;
(b)
in
thermal
reduced
conductivity
La4Ti4-xNbxO14+ ±δ;
of:
(c)
(a)
air
annealed
La4Ti4O14±δ;
(d)
La4Ti3.4Nb0.6O14.3±δ with temperature
Figure 6.14: Variation of thermal conductivity with Nb content for reduced
La4Ti4-xNbxO14+
±δ
at 573 K
Figure 6.15: Variation in thermal diffusivity of Sr4-xLaxNb4O14+
±δ
ceramics: (a) air
annealed; (b) reduced
Figure
6.16:
Variation
in
specific
heat
capacity
of:
(a)
Sr4Nb4O14±δ;
(b) Sr3.2La0.8Nb4O14.4±δ ceramics with temperature
Figure 6.17: Variation in thermal conductivity of: (a) air annealed Sr4-xLaxNb4O14±δ;
(b) reduced Sr4-xLaxNb4O14+
±δ;
(c) Sr4Nb4O14±δ; (d) Sr3.2La0.8Nb4O14.4±δ with
temperature
Figure 6.18: Variation of thermal conductivity with La content at 573 K: (a) air
annealed and; (b) reduced
Figure 6.19: Schematic representation of a typical non distorted La4Ti4O14 projected
along a axis
Figure 7.1: (a) X ray diffraction patterns for CoSb3 based materials, (b) SEM
micrograph of CoSb3
Page xiii
List of Figures and Tables
Figure 7.2: Variation of: (a) electrical resistivity; (b) Seebeck coefficient with
temperature for CoSb3 compositions
Figure 7.3: Variation of: (a) total thermal conductivity; (b) lattice thermal
conductivity with temperature for CoSb3 based compositions
Figure 7.4: Components of total thermal conductivity with T-1 for YCST
Figure 7.5: Temperature dependence of zT for CoSb3 based compositions
Page xiv
List of Figures and Tables
List of Tables
Table 4.1: Theoretical and relative densities of La4-xSrxTi4O14-
±δ ceramics
Table 4.2: amount of 5 layer PLS compound in La4-xSrxTi4O14Table 4.3: XPS data of La4-xSrxTi4O14-
±δ ceramics
±δ ceramics
Table 4.4: Theoretical and relative densities of La4Ti4-xTaxO14+
±δ
ceramics
Table 4.5: FWHM of the La4Ti3.6Ta0.4O14.2±δ ceramics
Table 4.6: XPS data of La4Ti3.6Ta0.4O14.2±δ ceramics
Table 4.7: Theoretical and relative densities of La4Ti4-xNbxO14+
±δ ceramics
Table 5.1: Theoretical and relative densities of Sr4-xLaxNb4O14+
±δ ceramics
Table 5.2: Amount of 2 layer PLS compound in Sr4-xLaxNb4O14+
±δ ceramics
Table 5.3: Percentage mass amount of different elements in Sr4-xLaxNb4O14+
±δ
ceramics
Table 5.4: XPS data of Sr3.2La0.4Nb4O14.2±δ ceramics
Table 6.1: Change in measured densities after Sr substitution and hydrogen
reduction
Table 6.2: Change in measured densities after Ta substitution and hydrogen
reduction
Table 6.3: Change in measured densities after Nb substitution and hydrogen
reduction
Table 6.4: Change in measured densities after La substitution and hydrogen
reduction
Table 6.5: Relative Formula weights, mean atomic volume and minimum thermal
conductivity of La4Ti4O14 and Sr4Nb4O14 based compositions
Page xv
List of Figures and Tables
Table 6.6: Phonon mean free path and Debye temperature for air annealed La4Ti4O14
and Sr4Nb4O14 based compositions
Page xvi
Chapter I. Introduction
Chapter I. Introduction
With increasing demand for energy, there is a need to develop new and
efficient ways to generate energy. Current efficiencies of internal combustion
engines and power generation plants are in the range of 30-38% with the remaining
energy being wasted into the environment as heat [1]. This heat can be utilized
usefully, for example, to generate electricity. Thermoelectric materials are one class
of environment friendly materials that can convert heat into electricity [2]. Unlike
traditional generators and engines, thermoelectric modules have no moving parts,
and are light weight with little to no service needed. These modules are not as
efficient (efficiency is <6%) as traditional generators and engines, but for smaller
applications, they are competitive with other technologies [3].
In order to establish the potential difference in a thermoelectric material, the
temperature gradient must be maintained which allows the electrons to flow from the
hot side to the cold side. For an ideal thermoelectric material, the thermal
conductivity must be on the lower side (typical values < 1W/m.K). Also, a high
electrical conductivity material (~ 8x105 S/m) is necessary to produced significant
flow of electrons [4]. The efficiency of a thermoelectric material depends upon
electrical conductivity (σ) and thermal conductivity (κ) and simultaneously on
Seebeck coefficient (S). These three quantities are linked together by a
dimensionless quantity zT.
zT=
(1.1)
Page 1
Chapter I. Introduction
One can increase either Seebeck coefficient or electrical conductivity or
lower thermal conductivity to make an efficient thermoelectric material. A zT value
of ~ 2 has been reported [5] which gives an efficiency of ~ 8.6% across a
temperature difference of 200 oC [6], which could be doubled if a zT value of 3 was
achievable across the same temperature range. As all of these properties are
interconnected, it is therefore difficult to improve the efficiency. This research
focuses on the effect of structural defects and compositional non stoichiometry on
the thermal conductivity of oxide and non-oxide materials.
Typical examples of thermoelectric materials are Bi2Te3, PbTe, skutterudites
(CoSb3, FeSb3 and ZnSb3), silicides (Mg2Si and MnSi1.72-1.74) and layered oxides
(Ca3Co4O9, Sr4Nb4O14, SrTiO3). Commercially used bismuth telluride (Bi2Te3)
operates near room temperature due to its low melting point and maximum
efficiency at this temperature [7]. Power generation industries involve applications
of thermoelectric materials at high temperatures ( > 800 oC) [8] and thermoelectric
materials with high operating temperatures are desirable. The focus of this research
will therefore be on skutterudites and oxides thermoelectric materials.
Skutterudites are efficient for intermediate temperature applications (25 oC to
~ 600 oC) [9, 10], such as cobalt antimonide (CoSb3) [11]. CoSb3 has a cubic
structure (space group Im3) with two naturally formed atomic cages per unit cell. It
has a melting point of 873 oC [12] but its high thermal conductivity (~ 10 W/m.K)
limits its zT value [13]. Skutterudites are good compounds for utilizing the idea of a
phonon glass electron crystal (PGEC), in which electrical properties are separated
from thermal properties and optimized at the same time [14]. Their cage like
structure favours the PGEC approach and can be partially filled with a variety of
atoms like rare earths or alkali metals [15]. These atoms are loosely bonded and
Page 2
Chapter I. Introduction
rattle inside the cage which causes phonons to scatter, which effectively decreases
thermal conductivity without degrading the electrical properties [16, 17] .
Oxide thermoelectric materials on the other hand possess higher melting
temperatures and they are chemically stable in air. The perovskite-like layer structure
(PLS) materials belongs to the homologous series of AnBnO3n+2 where n denotes the
number of BO6 octahedral layers. These materials inherently have lower thermal
conductivity values due to their layered structure. These layers scatter phonon more
efficiently giving low thermal conductivity compared to the perovskite materials
which are the end member of the homologous series AnBnO3n+2 (with n=∞). For
example, single crystal Sr4Nb4O14 has a thermal conductivity of ~ 2 W/m.K [18]
while SrTiO3 has a much higher thermal conductivity of ~ 12 W/m.K [19]. The
reason for anomalously low thermal conductivity for PLS compounds lies in the fact
that both PLS and perovskite compounds have the same thickness of the octahedral
layers, but PLS compounds have a layered crystal structure [20]. These atomic scale
layers help to reduce thermal conductivity by creating extra phonon scattering
centres.
Due to the unique crystal structure of skutterudite and PLS materials, it is
desirable to research these materials for their low thermal conductivity. This work
focuses on the investigation of skutterudites and PLS materials fabricated by Spark
plasma sintering. Thermal conductivity and microstructure of the materials were
studied in relation to the compositional non stoichiometry and oxidation-reduction.
Page 3
Chapter I. Introduction
REFERENCES
[1] K.M. Saqr, M.N. Musa, Therm Sci, 13 (2009) 165-174.
[2] G.F. Rinalde, L.E. Juanico, E. Taglialavore, S. Gortari, M.G. Molina,
International Journal of Hydrogen Energy, 35 (2010) 5818-5822.
[3] F.J. Weinberg, D.M. Rowe, G. Min, J Phys D Appl Phys, 35 (2002) L61-L63.
[4] G. Min, Journal of Electronic Materials, 39 (2010) 2459-2461.
[5] R. Venkatasubramanian, E. Siivola, T. Colpitts, B. O'Quinn, Nature, 413 (2001)
597-602.
[6] G. Chen, D. Kraemer, A. Muto, K. McEnaney, H.P. Feng, W.S. Liu, Q. Zhang,
B. Yu, Z.F. Ren, Micro- and Nanotechnology Sensors, Systems, and Applications Iii,
8031 (2011).
[7] O.C. Yelgel, G.P. Srivastava, Journal of Applied Physics, 113 (2013).
[8] R. Funahashi, S. Urata, Int J Appl Ceram Tec, 4 (2007) 297-307.
[9] G. Li, J.Y. Yang, Y. Xiao, L.W. Fu, J.Y. Peng, Y. Deng, P.W. Zhu, H.X. Yan,
Journal of Electronic Materials, 42 (2013) 675-678.
[10] J.J. Zhang, B. Xu, L.M. Wang, D.L. Yu, Z.Y. Liu, J.L. He, Y.J. Tianb, Applied
Physics Letters, 98 (2011).
[11] A. Harnwunggmoung, K. Kurosaki, H. Muta, S. Yamanaka, Applied Physics
Letters, 96 (2010) 202107.
[12] H.Y. Geng, S. Ochi, J.Q. Guo, Applied Physics Letters, 91 (2007).
[13] G.S. Nolas, M. Kaeser, R.T. Littleton, T.M. Tritt, Applied Physics Letters, 77
(2000) 1855.
Page 4
Chapter I. Introduction
[14] G.A. Slack, New Materials and Performance Limits for Thermoelectric Cooling,
in: D.M. Rowe (Ed.) CRC Handbook of Thermoelectrics, CRC, Boca Raton, 1995,
pp. 407.
[15] B.C. Sales, D. Mandrus, R.K. Williams, Science, 272 (1996) 1325-1328.
[16] X. Shi, H. Kong, C.P. Li, C. Uher, J. Yang, J.R. Salvador, H. Wang, L. Chen,
W. Zhang, Applied Physics Letters, 92 (2008) 182101.
[17] H. Li, X.F. Tang, Q.J. Zhang, C. Uher, Applied Physics Letters, 93 (2008).
[18] A. Sakai, T. Kanno, K. Takahashi, A. Omote, H. Adachi, Y. Yamada, X.D.
Zhou, Journal of the American Ceramic Society, 95 (2012) 1750-1755.
[19] H. Muta, K. Kurosaki, S. Yamanaka, Journal of Alloys and Compounds, 392
(2005) 306-309.
[20] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
Page 5
Chapter II. Literature Review
Chapter II. Literature Review
2.1
INTRODUCTION TO THERMOELECTRICS
Thermoelectrics are materials which have the ability to convert a temperature
gradient to an electromotive force by utilizing the Seebeck effect. Although this
effect was discovered in the earlier part of the 19th century [1], thermoelectric
generation was only practical in the second half of the 20th century [2]. The voltage
generated by metals is low (~ 50 µV/K) while semiconductors can generate much
higher voltages (several hundreds of µV/K). A thermoelectric module is a device
fabricated by using n type (in which electrons are the charge carriers) and p type (in
which holes are the charge carriers) elements to enhance the overall voltage. These
two elements are connected electrically in series and thermally in parallel as shown
in Figure 2.1.
Figure 2.1:
Typical thermoelectric modules: (a) for cooling; (b) for power generation
Page 6
Chapter II. Literature Review
2.2
EFFICIENCY OF THERMOELECTRIC MATERIALS
Generally the performance of a thermoelectric material is expressed in the
form a dimensionless figure of merit zT. ‘zT’ denotes the figure of merit of a
thermoelectric material, while ‘ZT’ denotes the figure of merit of a thermoelectric
module.
(2.1)
Where σ is electrical conductivity, S is the seebeck coefficient, κ is thermal
conductivity and T is absolute temperature. A high Seebeck coefficient, high
electrical conductivity and low thermal conductivity make for an efficient
thermoelectric material. zT values of more than 2 have been reported to date [3-5].
The efficiency of a thermoelectric material is given by [6]
√
√
(2.2)
ΔT is the temperature difference between the hot and the cold side. Th is the
temperature of the hot side (in Kelvin) and Tc is the temperature of the cold side.
Efficiencies of 5-10% have been reported so far [7].
Apart from the figure of merit, a more convenient term to use is power factor,
which is defined as S2σ. This term is useful because S and σ depend more strongly
on charge carrier concentrations while κ depends more strongly on the lattice
contribution to thermal conductivity [8]. However all of these properties are
Page 7
Chapter II. Literature Review
interconnected and it is therefore difficult to improve zT. Since electrical
conductivity and Seebeck coefficient are inversely proportional to each other, a
compromise has to be made between these two to obtain optimum values.
Semiconductors are the better choice for thermoelectric materials since they have
reasonable Seebeck coefficient and electrical conductivity values and can be further
tuned by doping to improve power factor values. Slack suggested the idea of the
phonon glass electron concept (PGEC) in which the electrical properties are
separated from thermal properties [9]. By utilizing the PGEC, thermal conductivity
can be reduced significantly without affecting electrical conductivity and the
Seebeck coefficient [10]. The material is ideally designed in such a way that it
behaves as a glass in terms of scattering phonons to give a low thermal conductivity.
At the same time the material has the electrical properties of a narrow band gap
semiconductor. Skutterudites are the best example to illustrate this concept which
will be discussed later.
Figure 2.2: (a) zT values of different thermoelectric materials [11]; (b) development in
thermoelectrics over the past 20 years [12]
Page 8
Chapter II. Literature Review
Until the 1990s, conventional thermoelectric materials were developed based
on the above mentioned characteristics. These materials tend to have heavy elements
for lower thermal conductivity and covalent bonding character for higher charge
mobility. Examples of these kinds of materials are Bi2Te3 [13] and PbTe [14].
Figure 2.2 shows typical thermoelectric materials and the improvement in zT with
time for thermoelectric materials. Bi2Te3 is the most widely used thermoelectric
material due to its high inherent Seebeck coefficient value of (200 µV/K) and
electrical conductivity (2x105 S/m) [15]. Great care must be taken during fabrication
of these materials as they readily oxidise when left in open air. The toxicity of these
materials is also a big issue and the powders of these materials must be kept under a
controlled environment.
2.3
THERMOELECTRIC MATERIALS
2.3.1 Skutterudites
Skutterudites have the general formula AX3 where A is a transition metal
such as Co or Rh and X is a pnicogen (a member of group 15 of periodic table) atom
such as As or Sb. Each transition metal forms a MX6 octahedra by coordinating with
6 pnicogen atoms. It has a cubic structure with 32 atoms per unit cubic cell as shown
in Figure 2.3.
Page 9
Chapter II. Literature Review
(a)
(b)
Figure 2.3: (a) crystal structure of typical skutterudite Rh4Sb12 filled with In (red balls), while blue
balls represent Rh atoms and yellow balls Sb atoms; (b) shows a cage formed by 12 Sb atoms with In
filler inside [16]
Skutterudites such as cobalt antimonide (CoSb3) [19] have been the focus of
research for intermediate temperature thermoelectric applications (298 K to 850 K)
[17, 18]. Caillat et al. first compared thermoelectric properties of single crystalline n
and p type CoSb3 grown by the gradient freeze technique [20]. They estimated the
effective mass of holes and electrons using hall mobility and Seebeck coefficient
measurements. The hall mobility of p type was much higher than the n type at room
temperature and their effect on Seebeck coefficient and electrical resistivity is shown
in the Figure 2.4.
Figure 2.4: Effect of hall carrier concentration on electrical resistivity and Seebeck coefficient [20]
Page 10
Chapter II. Literature Review
Despite their good power factor values, the thermal conductivity of
Skutterudites is on the higher side (~ 10 W/m.K) [10]. As mentioned earlier, the
crystal structure of skutterudites is excellent for applying the PGEC concept. It has
interstitial space for dopant elements that can be filled without changing the
structure. The doping elements drastically change the thermoelectric properties
especially thermal conductivity and increase zT values. The phonon mean free path
is kept as small as possible while the electron mean free path as long as possible
[21,22]. Sales et. al first reported Ce and La filled Fe4-xCoxSb12 by melt quenching
and hot pressing [23]. They reported a zT value of ~ 1 at 800 K with further
enhancement to 1.4 after optimizing the composition. This high zT value was the
result of low thermal conductivity caused by the ‘‘rattling’’ of La or Ce atoms inside
the cage like structure of FeSb3.
Chen et al. found that the 44 % of the cages in CoSb3 can be filled if Ba is
used as filler [24]. They used a two-step method for synthesising this composition. In
the first step, BaSb3 was synthesised by reacting Ba and Sb at 903 K and then in the
second step, additional Sb and Co were added and melted at 1323 K. Increasing Ba
concentration decreased the Seebeck coefficient, but increased electrical and thermal
conductivity which achieved a zT value of 1.1 at 850 K.
Indium and Ce doping into the cage like structure of CoSb3 gave a zT value
of 1.43 at 800 K [25]. The materials were prepared by melt spinning and park plasma
sintering. The Seebeck coefficient increased with increasing indium doping due to an
increased carrier concentration. Indium formed nanoscale InSb2 phase precipitates in
the matrix which was believed to enhance the Seebeck coefficient. For double filled
CoSb3 with Yb and Ba, synthesised by melt spinning and spark plasma sintering, a
zT value of ~ 1 at 773 K was reported [26]. They used different speeds of the
Page 11
Chapter II. Literature Review
quenching wheel which affected the individual thermoelectric properties, but the
overall zT remained almost the same. The minimum thermal conductivity obtained
was ~ 3.2 W/m.K, which is still higher for the doped CoSb3 materials. Double filled
Yb and Ba and Fe substituted CoSb3 synthesised by melting, annealing and hot
pressing had a thermal conductivity value of 1.8 W/m.K. This is amongst the lowest
values reported for skutterudite materials. The power factor on the other hand had a
lower value than Yb and Ba filled CoSb3 [26].
Li et al. reported a high zT value of 1.22 at 800 K in Yb filled CoSb3 [27].
The ingots of Yb0.2Co4Sb12 were prepared by melt quenching. These ingots were
inserted into a cylindrical quartz tube and melt spun at different speeds to obtain thin
ribbons which were then sintered by SPS. They concluded that higher speeds of melt
spinning increased the purity of the phase and decreased grain size. This affected the
thermal conductivity and also contributed in increasing the Seebeck coefficient,
while electrical conductivity seemed independent of the speed as shown in Figure
2.5.
Page 12
Chapter II. Literature Review
Figure 2.5: Variation of: (a) electrical conductivity; (b) Seebeck coefficient; (c) thermal
conductivity; (d) zT for Yb0.2Co4Sb12with temperature [27]
Changing the amount of Sb in Yb0.2CoSb3+y increased the zT slightly
reaching a value of 1.26 at 800 K for Yb0.2CoSb12.3 [28]. Air annealing had no effect
on overall zT [29], which suggests that materials melt spun in air are stable.
P type skutterudite, FeSb3 had a zT value of ~ 0.9 at 800 K when filled with
Ce and synthesised by melt spinning and SPS [30]. CeyFe8-xNixSb24 was prepared by
melt spinning, but its thermoelectric properties were not reported [31].
CoSb3 has a melting point of 873oC [32] with high carrier concentration and
Seebeck coefficient (~ 60 µV/K) [33]. However, due to strong covalent bonding, the
intrinsic thermal conductivity is ~ 10 W/m.K, which is a high value for
thermoelectric applications [34]. This high thermal conductivity limits the zT of
Page 13
Chapter II. Literature Review
CoSb3 based materials. Nano structuring is another way to reduce thermal
conductivity of CoSb3, but it also decreases its electrical conductivity by creating
more scattering centres for electrons [35]. In addition, making nanostructured
materials is not a cost effective processing route. In order to improve the electrical
properties of CoSb3, tellurium (Te) substitution for antimony (Sb) is effective [36].
A significant reduction in electrical resistivity for Co4Sb12-xTex was observed as
compared to pure CoSb3 because Te serves as an electron donor in the CoSb3
structure [37, 38].
CoSb3 has a cubic structure (Space group Im3) with two naturally formed
atomic cages per unit cell. These cages favour the PGEC approach and can be
partially filled with a variety of atoms like rare earths or alkali metals [23]. These
atoms are loosely bonded and rattle inside the cage, which causes phonons to scatter
and effectively decreases thermal conductivity without degrading the electrical
properties [39, 40]. Ytterbium (Yb) is an effective additive to the CoSb3 structure
due to its high atomic mass and small size. By partially filling the cages in CoSb 3
structure with Yb, the thermal conductivity is significantly decreased [41] while only
slightly affecting Seebeck coefficient and electrical resistivity [34]. Tellurium (Te)
substitution for antimony (Sb) is effective in improving the electrical properties of
CoSb3, [36]. A significant reduction in electrical resistivity for Co4Sb12-xTex was
observed as compared to pure CoSb3 because Te serves as an electron donor in the
CoSb3 structure [37, 38].
In terms of industrial applications, it is desirable for both the n and p type
legs of a thermoelectric module to have similar thermal and mechanical properties to
avoid any mechanical or thermal stress during thermal cycling. CoSb3 is an excellent
Page 14
Chapter II. Literature Review
choice for medium temperatures as both n and p type materials can be made from the
same material.
2.4
OXIDE THERMOELECTRICS
High processing cost, toxicity, low working temperatures (doped Bi2Te3
working temperature less than 200oC) and environmental hazards of traditional
thermoelectric materials are always a concern. This has stimulated the idea of using
oxides as thermoelectrics. Oxide materials on the other hand use nontoxic,
inexpensive and environmentally friendly elements. They are stable in air even at
higher temperatures (> 500oC). Oxide thermoelectrics generally have low Figure of
merit as compared to the bismuth telluride and skutterudites, but these materials have
no or less effect on environment and are chemically stable.
In the 1980s researchers started looking into oxide thermoelectrics. They first
investigated oxide superconductors, but their thermoelectric power values were only
in the order of a few µV/K [42, 43]. The superconductors did not appear to be
promising materials for thermoelectric application; however the layered cobalt
oxides showed some promising high thermoelectric performance [44, 45].
2.4.1 Layered Cobalt oxides and cobaltate
Layered cobalt oxides and cobaltate structures offer promising thermoelectric
properties due to phonon scattering at the interface between the layers giving low
thermal conductivity [44]. All of these oxides have a CdI2 type structure with CoO2
layers alternatively stacked with various NaCl type block layers (Figure 2.5), except
Page 15
Chapter II. Literature Review
for NaCo2O4 which has Na atoms inserted in between the CoO2 sheets. The most
commonly investigated compounds of the layered oxide thermoelectric materials are
NaxCoO2, Ca3Co4O9 and Ca2Co2O5 as shown in Figure 2.6.
Figure 2.6:
Crystal structures of some of the layered cobalt oxides [45]
Terasaki et al. reported the power factor of NaCo2O4 to be higher than Bi2Te3
at room temperature, which set the ground for oxide thermoelectric research [46]. In
general, the zT values for NaCo2O4 remained the same for doped and undoped
materials [47, 48]. On the other hand NaxCo2O4 has low humidity resistance and the
volatility of Na is a big issue in controlling its composition. In addition Na forms
electrically insulating compounds which decreases thermoelectric properties [12].
Ca3Co4O9 and Ca2Co2O5 are more commonly used due to their chemical and
compositional stability. Due to a strong anisotropy in the structure, single crystal
Ca3Co4O9 shows enhanced thermoelectric properties compared to the bulk
polycrystalline ceramic. Single crystal Ca3Co4O9 reported to have a zT value of 0.87
at 973 K [49], but for polycrystalline Ca3Co4O9 the zT value was ~ 0.3 at 1000 K
[50]. Sintered Ca3Co4O9 nanofibres densified by SPS were reported to have a zT
value of 0.16 at 800 K [51]. Partially doped Ca3Co4O9 with Bi was estimated to
Page 16
Chapter II. Literature Review
produce a zT value of 1.2 -2.7 for single crystal Ca2Co2O5 whiskers at temperatures
greater than 873 K [52].
2.4.2 ZnO based thermoelectrics
ZnO is a well-known wide band gap semiconductor which exists in two
crystallographic forms, hexagonal (wurtzite), cubic (zinc blende). Wurtzite is the
most stable under ambient conditions and shown in Figure 2.7. ZnO is n type
semiconductor with a band gap of ~ 3.3 eV at room temperature and a carrier
mobility of 200 cm2/Vs for the single crystal [53]. One disadvantage of ZnO is a
high thermal conductivity which is around 54 W/m.K for the single crystal [54] and
30 W/m.K. for the bulk ceramic at 300 K [55].
Figure 2.7:
Crystal Structure of ZnO (wurtzite) [56]
Al is the most commonly used dopant in ZnO. A small amount of Al
significantly increases the electrical conductivity [57]. Ohtaki et al. reported that
double doping of ZnO with Al and Ga achieved a zT value of 0.65 at 1247 K with a
Page 17
Chapter II. Literature Review
room temperature thermal conductivity of 5W/mK [58]. Attempts have been made to
decrease thermal conductivity, but it degrades the electrical conductivity and
Seebeck coefficient which in turn decreases zT. Al doped ZnO nanocomposite
showed a thermal conductivity of 2 W/m.K but produced a lower zT of 0.44 [59].
Similarly MgO doped Zn1-xAlxO suppressed the thermal conductivity to half the
value of undoped ZnO, but it decreased electrical conductivity which in turn was
unsuccessful in improving zT [60].
2.4.3 Indium based oxides
Indium oxide (In2O3) is highly electrically conductive and is used in many
industrial applications as a conductive coating. It has bixibyite type cubic crystal
structure (Figure 2.8). It has an electrical conductivity value of 0.34×105 S/m [61]
while its thermal conductivity value is ~ 10 W/m.K.
Figure 2.8:
Typical structure of In2O3 crystal
Page 18
Chapter II. Literature Review
Ge doping in In2O3 reduced its thermal conductivity from 10 W/m.K to 3
W/m.K, which in turn gave a zT value of 0.46 at 1273 K [62]. The thermal
conductivity was reduced by scattering of phonons from the fine precipitates of
In2Ge2O7. Ce doped nanostructured In2O3 was reported to have a zT value of 0.4 at
1050 K by reducing the thermal conductivity to 2.2 W/m.K, but sacrificing on
electrical conductivity [63]. SPS processing and Co doping produced ceramic with a
grain size of 0.3-0.4 µm. This resulted in a thermal conductivity value of 1.8 W/mK
for 80 % dense material, which is the lowest ever thermal conductivity reported for
an indium based oxide [64]. Despite very low thermal conductivity, the zT value was
0.26 at 1073 K due to a decrease in the Seebeck coefficient value.
Guilmeau et. al found that low doping levels of Ti, Zr, Sn, Ta and Nb (0.0020.006 mole percent) in In2O3 gives the highest zT [61] since higher doping levels
tend to increase thermal conductivity due to an increase in the electronic part of
thermal conductivity as shown in Figure 2.9.
Figure 2.9:
Influence of Sn and Ti in In2O3 on thermal conductivity and zT
Page 19
Chapter II. Literature Review
2.4.4 Perovskite
In 1839 Gustav Rose discovered the mineral Perovskite (CaTiO3) which was
named in the honour of Russian Mineralogist Lev Alexeievich Perovsky [65]. After
its discovery, many compounds with similar stoichiometric ratio and structure were
identified. Figure 2.10 represents a typical perovskite unit cell.
Figure 2.10: Cubic perovskite unit cell. Red spheres represent the A cation, Blue spheres represent
the B cation and Grey spheres represent X anions [66]
The structure of perovskite is described by a general formula ABX3, where A
and B are cations and X is an anion. Generally, the A cations are large and are
positioned at the corners of the cube while the B cations are small and positioned at
the centre of the unit cell in an octahedrally coordinated site. Both of the cations can
have a variety of charges. Oxygen is the anion and it occupies the face centred
position in the cube. The perovskite structure is more commonly represented in the
form of oxygen octahedra with B cations in the centre of octahedra and A cation in
between the oxygen octahedra. The octahedra are generally connected to each other
by the corner giving a typical perovskite structure as shown in Figure 2.11. For these
Page 20
Chapter II. Literature Review
reasons, the perovskite structure has great flexibility to accommodate different
elements and these octahedral layers can give rise to different layered compounds
with highly anisotropic properties. Examples of compounds with perovskite structure
are CaTiO3, CaMnO3 and BaTiO3.
Figure 2.11: Perovskite crystal structure showing the oxygen octahedra [67]
2.4.4.1 Strontium Titanate
SrTiO3 has a cubic perovskite structure at room temperature (space group
Pm3m) with a wide band gap of ~ 3.2 eV. It has applications in the field of
superconductivity, ferroelectricity and dielectric capacitors. In the pure form, SrTiO3
does not behave as a good thermoelectric material but doping, alloying and
introduction of oxygen vacancies can significantly enhance its thermoelectric
properties [68]. Kinaci et al. discussed the effect of dopants on the electronic
transport properties of SrTiO3 ceramic [69]. At higher concentration of dopants (12
Page 21
Chapter II. Literature Review
mol. %), lattice thermal conductivity is reduced due to an increase in scattering
centres, but the electronic part of the thermal conductivity also increased.
La doping increased the zT value of SrTiO3 by increasing its carrier
concentration and reached a value of ~ 0.37 at 1073 K for single crystal SrTiO3 [70]
and 0.21 for polycrystalline SrTiO3 [71]. The oxygen vacancies and La doping acted
as electron donors which increased the electrical conductivity. La doped SrTiO3
sintered using SPS produced a zT value of 0.37 at 1045 K, which is the highest value
for La doped SrTiO3 poly crystals [72].
Nb doping in SrTiO3 created larger carrier effective mass by expanding the
unit cell resulted in a higher Seebeck coefficient. A zT value of 0.27 at 1073K was
obtained when Nb was doped into single crystal SrTiO3 [70], while Nb doping of
epitaxial thin films gave a zT value of 0.37 [73]. A Seebeck coefficient value of 480
µV/K was reported for a single layer of Nb doped SrTiO3 due to strong confinement
of electrons. From this idea a two dimensional electron gas was fabricated at the
interface of a TiO2/SrTiO3 interface and a huge increase in Seebeck coefficient value
was observed. The value reached up to 1050 µV/K at room temperature. By taking
the minimum thermal conductivity of SrTiO3 based materials, a theoretical zT value
of ~ 2.4 was calculated [74].
2.4.4.2 Manganates
CaMnO3 has a typical perovskite structure as shown in Figure 2.12. It has a
room temperature electrical resistivity value of 0.3 Ωcm [75]. Substitution of Ca2+
ions by trivalent ions greatly enhances the transport properties [75] and the
octahedral layers help to reduce thermal conductivity by scattering more phonons.
Page 22
Chapter II. Literature Review
Figure 2.12: Typical structure of distorted CaMnO3 [76]
Formation of oxygen vacancies during fabrication process should be avoided
in n type material, because they degrade the electrical conductivity [77]. Ultrasonic
combustion synthesis of Nb doped CaMnO3 yielded very fine spheroidal particles
(10-50 nm) of manganate phase [78]. A zT value of 0.32 was reported at 1060 K. A
low thermal conductivity value of ~ 0.5 W/m.K was reported in 67-80% dense
samples, which can be related to the phonon scattering from the pores.
2.5
PEROVSKITE RELATED STRUCTURES
Layered perovskite structures have attracted attention due to their highly
anisotropic properties. By orienting the planes in one preferred direction, their
properties can be tuned to optimise their thermoelectric properties. There are many
Page 23
Chapter II. Literature Review
structures related to the perovskite structure. Examples of these types of structures
are Aurivillius [79], Dion Jacobson structure [80], Ruddlesden-Popper (RP) [81] and
Perovskite Like Layered Structure (PLS) [82].
Aurivillius phases are composed of perovskite blocks with general formula
An−1BnO3n+1 separated and shared by alternating Bi2O2 layers. A can be mono, di or
trivalent element (or combination), B is a transition element and n is the number of
octahedral layers in the perovskite structure [83]. The representative compounds
include Bi2WO6 (n=1), SrBi2Ta2O9 (n=2) and Bi4Ti3O12 (n=3) as shown in Figure
2.13.
Figure 2.13: Schematic drawing of structure of Aurivillius compounds Bi2WO6 (n=1), SrBi2Ta2O9
(n=2) and Bi4Ti3O12 (n=3) [83]
Dion-Jacobson type phases (A’An-1BnO3n+1), perovskite-like layered structure
(PLS) (AnBnO3n+2) and hexagonal type phases AnBn-1O3n are related structures and
they can be considered by cutting the cubic perovskite ABO3 structure along [100],
[110] and [111] direction, respectively and inserting an additional row of oxygen.
The respective structure is given in the Figure 2.14.
Page 24
Chapter II. Literature Review
Figure 2.14: Crystal Structure of Dian Jacobson phase CsBiNb2O7 [84]
Ruddlesden-Popper (RP) phases are another type of layered perovskites,
They have the general formula An+1BnO3n+1 (or AO(ABO3)n) where B is Ti, Mn, Al
or Nb. There crystal structure generally consists of alternating perovskite type layer
of ABO3 located within the rock salt type layers of AO. Corner sharing BO6
octahedra form layers, with A cations occupying the interstitial sites with 9 and 12
coordination. The first characterised RP phases were the SrO(SrTiO3)n series by
Ruddlesden and Popper [85] and shown in Figure 2.15.
Page 25
Chapter II. Literature Review
Figure 2.15: Crystal structure of Sr3Ti2O7 [86]
AnBnO3n+2 (PLS) compounds are formed by the perovskite layers when they
are cut through the (110) plane of the cubic perovskite structure. n in the general
formula represents the number of BO6 octahedra (or layers) in one block. In the case
of a mixed number of layers, n specifies the average number of octahedral layers per
block. The A cations are often alkaline earth or lanthanide metals and the B cations
are usually titanium, tantalum or niobium. The general formula for all the PLS
compound is same but they have different structure and properties due to different
numbers of layers, different tilting of BO6 octahedra and displacement of A and B
cations [87].
Figure 2.16 shows some of the common PLS compounds with different
numbers of layers (n =2 to ∞).
Page 26
Chapter II. Literature Review
Figure 2.16: Structural diagram of non-distorted AnBnO3n+2 projected along a axis
2.5.1 Properties of Layered Perovskite Thermoelectrics
Layered perovskites have been studied extensively for their ferroelectric and
piezoelectric properties. These materials are known for their high curie points. The
Curie temperature is defined as the temperature at which a material experiences a
phase transition from (low-temperature) ferroelectric to (high-temperature)
paraelectric phase. La4Ti4O14 has a curie point of 1485 oC [88] while Sr4Nb4O14 has
1325 oC [82]. As mentioned before, SrTiO3 doped ceramics tend to have high thermo
power values in the range of -550 to -700 µV/K [89, 90]. However, due to their
relatively high thermal conductivity (~ 10 W/m.K), their zT value are very low. In
layered perovskite structure ceramics, for instance the RP structure, they tend to have
a rock salt type layer in between the perovskite structure. This rock salt layer serves
as a phonon scattering centre, which decreases the phonon mean free path and as a
result lowers its thermal conductivity. These kinds of structural features are common
Page 27
Chapter II. Literature Review
in layered perovskite structures which make them available to use in variety of
applications for example in thermal barrier coatings [91, 92]. By tuning the
stoichiometry, thermoelectric properties of layered perovskites can be enhanced [93].
2.5.1.1 Aurivillius Structure Thermoelectrics
Bi4Ti3O12 is a member of the aurivillius family with n= 3. Its structure can be
defined by alternate stacking of [Bi2O2]2+ and [Bi2Ti3O10]2- layers. Randomly
oriented ceramics of Bi4Ti3O12 have a thermal conductivity value of ~ 1.5 W/m.K
which decreased to ~ 1W/m.K when textured and shown in Figure 2.17 [94]. This
tells that the thermal conductivity is anisotropic in these materials and depends on
the orientation. The reason for such a low thermal conductivity is the phonon mean
free path being equal to the interatomic spacing along the pressing direction and it
cannot be reduced further which also explains the nearly temperature independent
nature of thermal conductivity [94].
Figure 2.17: Thermal conductivity as a function of temperature for Bi4Ti3O12 [94]
Page 28
Chapter II. Literature Review
2.5.1.2 Ruddlesden Popper (R-P) Structure Thermoelectrics
Srn+1TinO3n+1 with n = 2 is the most studied thermoelectric material in the RP
family. This structure has alternate stacking of NaCl type SrO and a perovskite type
SrTiO3 layer. This structure combines the excellent carrier transport features of
SrTiO3 with phonon scattering at the interface. Sr3Ti2O7 structures exhibit
anisotropic properties and showed a 50% drop in thermal conductivity at room
temperature when doped with Nb as compared to Nb doped SrTiO3 along the longest
axis [86]. It is observed that the electrical conduction behaviour of Nb and rare earth
doped Sr3Ti2O7 is almost the same. The electrical conductivity decreased with
increasing temperature for all compositions. A zT value of 0.25 was obtained for Gd
doped Sr3Ti2O7 as shown in Figure 2.18.
Figure 2.18: Temperature dependence of power factor, thermal conductivity and figure of merit
for rare earth doped Sr3Ti2O7 [86]
Page 29
Chapter II. Literature Review
Double substitution of Gd and Ta in Sr3Ti2O7 resulted in a lower thermal
conductivity value (~ 3.4 W/m.K), but the Seebeck coefficient dropped [95]. This
was due to the decrease in the effective mass of the charge carriers. Double
substitution of Ca and Nb in Sr3Ti2O7 gave a thermal conductivity value of ~ 4
W/m.K which decreased with increasing temperature [96]. The Seebeck coefficient
also dropped as the amount of dopant increased showing similar trends to Gd and Ta
substitution in Sr3Ti2O7 giving a zT value of ~ 1.5 at 1000 K. Single substitution of
Nb or La resulted in a zT value of 0.15 at 1000 K [97, 98]. This increase in zT was
due to the increased electrical conductivity and Seebeck coefficient values. These
results suggest single doping to be more effective than double doping.
2.5.1.3 Dion Jacobson Structure Thermoelectrics
Epitaxial layers of CsBiNb2O7 showed an extremely low thermal
conductivity of 0.4 W/m.K, when they were synthesised by pulsed laser deposition
technique. The reason for this ultra-low thermal conductivity is unknown, but the
author relates it to the large amount of defects in the crystal structure [99]. These
results may not be accurate due to the difficulty in physical property measurement of
thin films which incorporates large errors. Figure 2.19 compares thermal
conductivity of CsBiNb2O7 with other materials.
Page 30
Chapter II. Literature Review
Figure 2.19: Temperature dependence of thermal conductivity of various oxides [47]
2.5.1.4 A4B4O14 Thermoelectrics
A4B4O14 compounds belong to the PLS compounds (AnBnO3n+2) with n=4 i.e.
the unit cell contains 4 octahedral layers. Their molecular formula is A4B4O14, but
they are best known by their empirical formula A2B2O7. Examples of these structures
are Sr4Nb4O14, La4Ti4O14 and Ca4Nb4O14 [82, 100, 101].
Sr4Nb4O14 is also one of the derivatives of the layered perovskite structures
with general formula of SrnNbnO3n+2 (n=4) n denotes the number of corner shared
Page 31
Chapter II. Literature Review
NbO6 octahedra layers as shown in Figure 2.20. It has orthorhombic crystal system
and space group of Cmc21. The lattice parameters for Sr4Nb4O14 are a=3.933 Å,
b=26.726 Å, c=5.683 Å [102].
Figure 2.20: Crystal Structure of Sr4Nb4O14 [93]
Sr4Nb4O14 is a wide band gap semiconductor which makes it inherently
insulator but 1 mol % doping of La in a single crystal makes it electrically
conductive [93]. The single crystal was prepared by the floating zone method under a
mixture of Ar+0.75%H2. The room temperature resistivity across the b axis is high
as compared to the other two axes due to the scattering of electrons from NbO6
octahedral layers. Thermal conductivity also behaved in the same way and a very
low thermal conductivity value of 0.4 W/m.K was observed along the b axis as
shown in Figure 2.21. Similar results were obtained by Kobayashi et al. [103].
Page 32
Chapter II. Literature Review
Figure 2.21: Variation in: (a) thermal conductivity; (b) Seebeck coefficient; (c) electrical
resistivity with temperature for Sr3.6La0.4Nb4O14.2±δ [94]
When more La was added to the Sr4Nb4O14 the thermal conductivity dropped
further [104]. The measurement was done for all the axes but La doped Sr4Nb4O14
showed better properties along the c axis. The trend in thermal conductivity with
increasing La content along c axis is shown in the Figure 2.22.
Figure 2.22: Variation of room temperature thermal conductivity with increasing amount of La
[104]
Page 33
Chapter II. Literature Review
In the case of polycrystalline ceramics, prepared by the co precipitation
method, the thermal conductivity remained almost temperature independent [105].
Textured ceramics had lower values of thermal conductivity along the pressing
direction as compared to the randomly oriented ceramics as shown in Figure 2.23.
This was due to the phonon mean free path approaching the distance between
perovskite blocks and the length of individual perovskite blocks.
Figure 2.23: Thermal conductivity as a function of temperature for the textured and randomly
oriented polycrystalline Sr3.9La0.1Nb4O14 [105]
The room temperature crystal structure of La4Ti4O14 has been observed to be
monoclinic. The lattice parameters are a = 7.80 Å, b = 13.011 Å and c = 5.546 Å and
β = 98.6o with space group P21 [105]. At 780 oC the monoclinic structure transforms
into an orthorhombic structure and the lattice parameters change to a = 3.954 Å, b =
25.952 Å and c = 5.607 Å with space group Cmc21. PLS La4Ti4O14 doped with Sr
Page 34
Chapter II. Literature Review
showed an electrical conductivity value of 3.4x10-6 S/cm at 700 oC [107]. Calculated
minimum thermal conductivity value was found to be 1.46 W/m.K [106] but the
measured value was 2.28 W/m.K at room temperature [108].
Apart from PLS, La4Ti4O14 also exists in pyrochlore form which has cubic
crystal structure as shown in Figure 2.24.
Figure 2.24: Crystal structure of pyrochlore La2Ti2O7 [106]
2.6
THERMAL CONDUCTIVITY
Thermal conductivity is the ability of a material to conduct heat. In terms of
thermoelectrics, thermal conductivity plays a very important role. For efficient
thermoelectric materials, the temperature gradient must be maintained for an infinite
period of time. Unfortunately, every material has a minimum thermal conductivity,
due to the electrons and the lattice vibrations (phonons contributions). Researchers
Page 35
Chapter II. Literature Review
are trying different ways to reduce thermal conductivity by various methods to
improve zT.
Thermal conductivity is defined as when a temperature gradient is applied,
the heat flow rate through a unit length of a material in a direction perpendicular to a
unit area
⃗
⃗
(2.3)
Where L is a unit length (or thickness) Q is the heat flow rate, A is the cross
sectional area and T is the absolute temperature.
Thermal conductivity mainly comes from two contributions lattice and
electronic thermal conductivity. In materials with low electrical conductivity
(insulators) the contribution of total thermal conductivity mainly comes from lattice
vibrations. In the case of electrical conductors, the contribution to the total thermal
conductivity comes from electronic thermal conductivity and lattice thermal
conductivity. The lattice thermal conductivity can be calculated by subtracting the
electronic contribution to thermal conductivity (κelec) from total thermal
conductivity. κelec can be calculated by using Weidman- Franz Law which is written
as
κelec = LTσ
(2.4)
Where L is the Lorenz number, T is the absolute temperature and σ is the
electrical conductivity. Assuming the case of a homogeneous material with a
Page 36
Chapter II. Literature Review
parabolic band dominated by acoustic phonon scattering at low temperatures
(κLattice α T-1), the Lorenz number is given as [109]
2
2
 k   3F  F2    4 F1   
L   B   0
2

F0  
 e  

(2.5)
Where kB is the Boltzmann constant, e is the unit charge and Fn(  ) is the nth
order Fermi integral and  is the reduced Fermi energy which can be calculated by
considering the Fermi Dirac statistics, the Seebeck coefficient values can be
expressed as [109]
S
kB
e
 F1  

 2
  
 F0  

(2.6)
This holds true for the materials which have only one type of charge carrier. For
materials which have both types of charge carriers, i.e. electrons and holes, the total
electronic thermal conductivity can be represented by [8]
(2.7)
Where κe1 and κe2 are the partial thermal conductivity contributions by holes and
electrons respectively. Similarly σ1 and σ2 are the partial electrical conductivity
contributions by holes and electrons while S1-S2 is the partial difference in Seebeck
coefficient of both charge carriers. The third term [
] in the above
equation is related to bipolar diffusion. This happens most commonly in small
Page 37
Chapter II. Literature Review
energy gap semiconductors. Electrons-holes pairs are generated by the absorption of
energy at the hot end. These pairs recombine by releasing energy when they move to
the cold end and thus increasing thermal conductivity. This co-existence of both
charge carriers explains the increase in thermal conductivity with increasing
temperature. In compounds like CoSb3, the intrinsic contribution starts to operate at
higher temperature (> 400 oC) which causes the increase in thermal conductivity as
shown in the Figure 2.25 [37].
Figure 2.25: Temperature dependence of thermal conductivity in CoSb3-xTex [37]
2.7
OXIDE THERMOELECTRIC MODULES
A thermoelectric module fabricated by using p type Ca3Co4O9 and n type
(ZnO)7In2O3 legs generated a maximum power output of 423 mW with 44 p-n
junctions at a temperature difference of ~ 850 K (cold side at 427 K and hot side at
1100 K) [110]. The P type leg has a Z value of 0.55x10-4/K while the n type has a
value of 1.35 x10-4/K at 1100K. [110] .The main degrading factors for the low power
Page 38
Chapter II. Literature Review
generation are dry joints and pores originating from the difference between the
electrodes and legs of the module. Funahashi et al. also fabricated an oxide
thermoelectric module by using 140 pairs of these layered oxides (p-type
Ca2.7Bi0.3Co4O9 and n-type La0.9Bi0.1NiO3) and showed that a battery of a cell phone
can be charged with this [111]. A maximum power output of ~0.15 W was achieved
at a temperature difference of ~ 550 K. The contact resistance in the module can
effectively be reduced by mixing n or p type powder in the Ag paste used as the
adhesive material between thermoelectric legs and the substrate. A maximum power
of 0.17 W was achieved by using this technique in a module prepared using eight
pairs of p-type Ca2.7Bi0.3Co4O and n-type CaMn0.98Mo0.02O3 as shown in Figure 2.26
[112].
Figure 2.26: A typical oxide thermoelectric module made from Ca2.7Bi0.3Co4O9 and
CaMn0.98Mo0.02O3 [112]
Page 39
Chapter II. Literature Review
REFERENCES
[1] H.J. Goldsmid, The thermoelectric and related effects, in: J.P. Robert Hull, R.M.
Osgood, Jr., Hans Warlimont (Ed.) Introduction to thermoelectricity, Springer
Heidelberg Dordrecht London, New York, 2010, pp. 1.
[2] H.J. Goldsmid, R.W. Douglas, British Journal of Applied Physics, 5 (1954) 386.
[3] K. Biswas, J.Q. He, I.D. Blum, C.I. Wu, T.P. Hogan, D.N. Seidman, V.P. Dravid,
M.G. Kanatzidis, Nature, 489 (2012) 414-418.
[4] R. Venkatasubramanian, E. Siivola, T. Colpitts, B. O'Quinn, Nature, 413 (2001)
597-602.
[5] K.F. Hsu, S. Loo, F. Guo, W. Chen, J.S. Dyck, C. Uher, T. Hogan, E.K.
Polychroniadis, M.G. Kanatzidis, Science, 303 (2004) 818-821.
[6] G.J. Snyder, E.S. Toberer, Nature materials, 7 (2008) 105-114.
[7] A. Patyk, Appl Energ, 102 (2013) 1448-1457.
[8] G.S.N.a.H.J. Goldsmid, Thermal Conductivity of Semiconductors, in: T.M. Tritt
(Ed.) Thermal conductivity Theory, Properties and applications, Kluwer Academic /
Plenum Publishers, 2004, pp. 112.
[9] G.A. Slack, New Materials and Performance Limits for Thermoelectric Cooling,
in: D.M. Rowe (Ed.) CRC Handbook of Thermoelectrics, CRC, Boca Raton, 1995,
pp. 407.
[10] J. Khaliq, Q. Jiang, J. Yang, K. Simpson, H. Yan, M.J. Reece, Scripta
Materialia, 72–73 (2014) 63-66.
[11] J.-F. Li, W.-S. Liu, L.-D. Zhao, M. Zhou, NPG Asia Materials, 2 (2010) 152158.
[12] M. Ohtaki, Journal of Ceramic Society Japan, 119 (2011).
Page 40
Chapter II. Literature Review
[13] S. Grasso, N. Tsujii, Q.H. Jiang, J. Khaliq, S. Maruyama, M. Miranda, K.
Simpson, T. Mori, M.J. Reece, Journal of Materials Chemistry C, 1 (2013) 23622367.
[14] Z. Dashevsky, S. Shusterman, M.P. Dariel, I. Drabkin, Journal of Applied
Physics, 92 (2002) 1425.
[15] M. Saleemi, M.S. Toprak, S. Li, M. Johnsson, M. Muhammed, Journal of
Materials Chemistry, 22 (2012) 725.
[16] J. Eilertsen, J. Li, S. Rouvimov, M.A. Subramanian, Journal of Alloys and
Compounds, 509 (2011) 6289-6295.
[17] G. Li, J.Y. Yang, Y. Xiao, L.W. Fu, J.Y. Peng, Y. Deng, P.W. Zhu, H.X. Yan,
Journal of Electronic Materials, 42 (2013) 675-678.
[18] J.J. Zhang, B. Xu, L.M. Wang, D.L. Yu, Z.Y. Liu, J.L. He, Y.J. Tianb, Applied
Physics Letters, 98 (2011) 072109.
[19] A. Harnwunggmoung, K. Kurosaki, H. Muta, S. Yamanaka, Applied Physics
Letters, 96 (2010) 202107.
[20] T. Caillat, A. Borshchevsky, J.P. Fleurial, Journal of Applied Physics, 80 (1996)
4442-4449.
[21] S.W. Kim, Y. Kimura, Y. Mishima, Advanced Materials for Energy Conversion
Ii, (2004) 377-384.
[22] G.S. Nolas, C.A. Kendziora, Physical Review B, 59 (1999) 6189-6192.
[23] B.C. Sales, D. Mandrus, R.K. Williams, Science, 272 (1996) 1325-1328.
[24] L.D. Chen, T. Kawahara, X.F. Tang, T. Goto, T. Hirai, J.S. Dyck, W. Chen, C.
Uher, Journal of Applied Physics, 90 (2001) 1864-1868.
[25] H. Li, X.F. Tang, Q. J. Zhang, C. Uher, Applied Physics Letters, 94 (2009)
102114.
Page 41
Chapter II. Literature Review
[26] J.R. Salvador, R.A. Waldo, C.A. Wong, M. Tessema, D.N. Brown, D.J. Miller,
H. Wang, A.A. Wereszczak, W. Cai, Materials Science and Engineering: B, 178
(2013) 1087-1096.
[27] H. Li, X.F. Tang, X.L. Su, Q.J. Zhang, C. Uher, J Phys D Appl Phys, 42 (2009)
145409.
[28] H. Li, X. Tang, X. Su, Q. Zhang, Applied Physics Letters, 92 (2008) 202114.
[29] H. Li, X. Tang, Q. Zhang, Journal of Electronic Materials, 38 (2009) 12241228.
[30] G. Tan, W. Liu, S. Wang, Y. Yan, H. Li, X. Tang, C. Uher, J Mater Chem A, 1
(2013) 12657-12668.
[31] T. Morimura, M. Hasaka, Scripta Materialia, 48 (2003) 495-500.
[33] W.-S. Liu, B.-P. Zhang, J.-F. Li, L.-D. Zhao, Journal of Physics D: Applied
Physics, 40 (2007) 566-572.
[34] G.S. Nolas, M. Kaeser, R.T. Littleton, T.M. Tritt, Applied Physics Letters, 77
(2000) 1855.
[35] P.-X. Lu, F. Wu, H.-L. Han, Q. Wang, Z.-G. Shen, X. Hu, Journal of Alloys and
Compounds, 505 (2010) 255-258.
[36] X.Y. Li, L.D. Chen, J.F. Fan, W.B. Zhang, T. Kawahara, T. Hirai, Journal of
Applied Physics, 98 (2005).
[37] W.-S. Liu, B.-P. Zhang, J.-F. Li, H.-L. Zhang, L.-D. Zhao, Journal of Applied
Physics, 102 (2007) 103717.
[38] M.J. Kim, I.H. Kim, Met Mater Int, 16 (2010) 459-463.
[39] X. Shi, H. Kong, C.P. Li, C. Uher, J. Yang, J.R. Salvador, H. Wang, L. Chen,
W. Zhang, Applied Physics Letters, 92 (2008) 182101.
[40] H. Li, X.F. Tang, Q.J. Zhang, C. Uher, Applied Physics Letters, 93 (2008).
Page 42
Chapter II. Literature Review
[41] G. S. Nolas M. Kaeser, R. T. Littleton IV, T. M. Tritt, Applied Physics Letters,
77 (2000) 1855.
[42] M. Ausloos, K. Durczewski, S.K. Patapis, C. Laurent, H.W. Vanderschueren,
Solid State Communications, 65 (1988) 365-368.
[43] A.T. Burkov, A. Heinrich, M.V. Vedernikov, AIP Conference Proceedings, 316
(1994) 76-80.
[44] A. Chernatynskiy, R.W. Grimes, M.A. Zurbuchen, D.R. Clarke, S.R. Phillpot,
Applied Physics Letters, 95 (2009) 161906.
[45] I. Terasaki, M. Iwakawa, T. Nakano, A. Tsukuda, W. Kobayashi, Dalton Trans,
39 (2010) 1005-1011.
[46] I. Terasaki, Y. Sasago, K. Uchinokura, Physical Review B, 56 (1997) 1268512687.
[47] M. Ito, D. Furumoto, Journal of Alloys and Compounds, 450 (2008) 517-520.
[48] P.H. Tsai, T. Norby, T.T. Tan, R. Donelson, Z.D. Chen, S. Li, Applied Physics
Letters, 96 (2010) 141905.
[49] M. Shikano, R. Funahashi, Applied Physics Letters, 82 (2003) 1851-1853.
[50] D. Wang, L. Chen, Q. Yao, J. Li, Solid State Communications, 129 (2004) 615618.
[51] T. Yin, D. Liu, Y. Ou, F. Ma, S. Xie, J.-F. Li, J. Li, The Journal of Physical
Chemistry C, 114 (2010) 10061-10065.
[52] R. Funahashi, I. Matsubara, H. Ikuta, T. Takeuchi, U. Mizutani, S. Sodeoka,
Japanese Journal of Applied Physics 2, 39 (2000) L1127-L1129.
[53] P. Wagner, R. Helbig, Journal of Physics and Chemistry of Solids, 35 (1974)
327-335.
[54] G.A. Slack, Physical Review B, 6 (1972) 3791-3800.
Page 43
Chapter II. Literature Review
[55] K.F. Cai, E. Müller, C. Drašar, A. Mrotzek, Materials Science and Engineering:
B, 104 (2003) 45-48.
[56] Z.L. Wang, Materials Science and Engineering: R: Reports, 64 (2009) 33-71.
[57] M. Ohtaki, T. Tsubota, K. Eguchi, H. Arai, Journal of Applied Physics, 79
(1996) 1816-1818.
[58] M. Ohtaki, K. Araki, K. Yamamoto, Journal of Electronic Materials, 38 (2009)
1234-1238.
[59] P. Jood, R.J. Mehta, Y.L. Zhang, G. Peleckis, X.L. Wang, R.W. Siegel, T.
Borca-Tasciuc, S.X. Dou, G. Ramanath, Nano Letters, 11 (2011) 4337-4342.
[60] T. Tsubota, M. Ohtaki, K. Eguchi, H. Arai, Journal of Materials Chemistry, 8
(1998) 409-412.
[61] E. Guilmeau, D. Berardan, C. Simon, A. Maignan, B. Raveau, D.O. Ovono, F.
Delorme, Journal of Applied Physics, 106 (2009) 053715-053717.
[62] D. Berardan, E. Guilmeau, A. Maignan, B. Raveau, Solid State
Communications, 146 (2008) 97-101.
[63] J.L. Lan, Y.H. Lin, Y. Liu, S.L. Xu, C.W. Nan, Journal of the American
Ceramic Society, 95 (2012) 2465-2469.
[64] Y. Liu, Y.H. Lin, J.L. Lan, W. Xu, B.P. Zhang, C.W. Nan, H.M. Zhu, Journal of
the American Ceramic Society, 93 (2010) 2938-2941.
[65] M.C. Knapp, in Investigations into the structure and properties of ordered
perovskites, layered perovskites, and defect pyrochlores, The Ohio State University,
2006.
[66] E. Fabbri, D. Pergolesi, E. Traversa, Chemical Society Reviews, 39 (2010)
4355-4369.
Page 44
Chapter II. Literature Review
[67] S. Stolen, E. Bakken, C.E. Mohn, Physical Chemistry Chemical Physics, 8
(2006) 429-447.
[68] T. Okuda, K. Nakanishi, S. Miyasaka, Y. Tokura, Physical Review B, 63 (2001)
113104.
[69] A. Kinaci, C. Sevik, T. Çağın, Physical Review B, 82 (2010) 155114.
[70] S. Ohta, T. Nomura, H. Ohta, K. Koumoto, Journal of Applied Physics, 97
(2005) 092108.
[71] J. Liu, C.L. Wang, W.B. Su, H.C. Wang, P. Zheng, J.C. Li, J.L. Zhang, L.M.
Mei, Applied Physics Letters, 95 (2009) 162110.
[72] A. Kikuchi, N. Okinaka, T. Akiyama, Scripta Materialia, 63 (2010) 407-410.
[73] S. Ohta, T. Nomura, H. Ohta, M. Hirano, H. Hosono, K. Koumoto, Applied
Physics Letters, 87 (2005) 092108.
[74] H. Ohta, S. Kim, Y. Mune, T. Mizoguchi, K. Nomura, S. Ohta, T. Nomura, Y.
Nakanishi, Y. Ikuhara, M. Hirano, H. Hosono, K. Koumoto, Nature Materials, 6
(2007) 129-134.
[75] A. Srivastava, N.K. Gaur, Journal of Magnetism and Magnetic Materials, 321
(2009) 3854-3865.
[76] T. Okuda, Y. Fujii, Journal of Applied Physics, 108 (2010) 103702.
[77] F.P. Zhang, Q.M. Lu, X. Zhang, J.X. Zhang, Journal of Alloys and Compounds,
509 (2011) 542-545.
[78] L. Bocher, M.H.Aguire, D. Logvinovich, A. Shkabko, R. Robert, M. Trottmann,
and A. Weidenkaff, Inorganic Chemistry, 47 (2008) 8077.
[79] H.X. Yan, H.T. Zhang, Z. Zhang, R. Ubic, M.J. Reece, Journal of the European
Ceramic Society, 26 (2006) 2785-2792.
Page 45
Chapter II. Literature Review
[80] L. Viciu, V.O. Golub, J.B. Wiley, Journal of Solid State Chemistry, 175 (2003)
88-93.
[81] K.H. Lee, S.W. Kim, H. Ohta, K. Koumoto, Journal of Applied Physics, 100
(2006) 063717.
[82] H. Ning, H. Yan, M.J. Reece, Journal of the American Ceramic Society, 93 [5]
(2010) 1409.
[83] R. Machado, M.G. Stachiotti, R.L. Migoni, A.H. Tera, Physical Review B, 70
(2004) 214112.
[84] C.J. Fennie, K.M. Rabe, Applied Physics Letters, 88 (2006) 262902.
[85] F. Lichtenberg, A. Herrnberger, K. Wiedenmann, Progress in Solid State
Chemistry, 36 (2008) 253-387.
[86] Y. Wang, K.H. Lee, H. Hyuga, H. Kita, H. Ohta, K. Koumoto, Journal of
Electroceramics, 24 (2010) 76-82.
[87] I. Levin, L.A. Bendersky, Acta Crystallographica Section B-Structural Science,
55 (1999) 853-866.
[88] H. Yan, H. Ning, Y. Kan, P. Wang, M.J. Reece, Journal of the American
Ceramic Society, 92 (2009) 2270-2275.
[89] H. Frederikse, W. Thurber, W. Hosler, Physical Review, 134 (1964) A442A445.
[90] B. Jalan, S. Stemmer, Applied Physics Letters, 97 (2010) 042106.
[91] W. Ma, D.E. Mack, R. Vassen, D. Stover, Journal of the American Ceramic
Society, 91 (2008) 2630-2635.
[92] W. Ma, M.O. Jarligo, D.E. Mack, D. Pitzer, J. Malzbender, R. Vassyen, D.
Stover, Journal of Thermal Spray Technology, 17 (2008) 831-837.
Page 46
Chapter II. Literature Review
[93] A. Sakai, T. Kanno, K. Takahashi, Y. Yamada, H. Adachi, Journal of Applied
Physics, 108 (2010) 103706.
[94] Y. Shen, D.R. Clarke, P.A. Fuierer, Applied Physics Letters, 93 (2008) 102907.
[95] R.R. Sun, X.Y. Qin, L.L. Li, D. Li, J. Zhang, Q.Q. Wang, Journal of Physics D
Applied Physics, 45 (2012) 124904.
[96] K.H. Lee, S.W. Kim, H. Ohta, K. Koumoto, Journal of Applied Physics, 101
(2007) 326.
[97] R.R. Sun, X.Y. Qin, L.L. Li, D. Li, N.N. Wang, J. Zhang, Q.Q. Wang, Journal
of Applied Physics, 112 (2012) 124904.
[98] K.H. Lee, Y.F. Wang, S.W. Kim, H. Ohta, K. Koumoto, International Journal of
Applied Ceramic Technology, 4 (2007) 326-331.
[99] D.G. Cahill, A. Melville, D.G. Schlom, M.A. Zurbuchen, Applied Physics
Letters, 96 (2010) 121903.
[100] Z. Gao, H. Ning, C. Chen, R. Wilson, B. Shi, H. Ye, H. Yan, M.J. Reece, J.L.
Jones, Journal of the American Ceramic Society, 96 (2013) 1163-1170.
[101] Z.P. Gao, H.X. Yan, H.P. Ning, R. Wilson, X.Y. Wei, B. Shi, H. Ye, M.J.
Reece, Journal of the European Ceramic Society, 33 (2013) 1001-1008.
[102] N. Ishizawa, F. Marumo, T. Kawamura, M. Kimura, Acta Crystallographica
Section B, 31 (1975) 1912-1915.
[103] W. Kobayashi, Y. Hayashi, M. Matsushita, Y. Yamamoto, I. Terasaki, A.
Nakao, H. Nakao, Y. Murakami, Y. Moritomo, H. Yamauchi, M. Karppinen,
Physical Review B, 84 (2011) 085118.
[104] A. Sakai, T. Kanno, K. Takahashi, A. Omote, H. Adachi, Y. Yamada, X.D.
Zhou, Journal of the American Ceramic Society, 95 (2012) 1750-1755.
Page 47
Chapter II. Literature Review
[105] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
[106] B. Liu, J.Y. Wang, F.Z. Li, Y.C. Zhou, Acta Materialia, 58 (2010) 4369-4377.
[107] K.E. H. Takamura, A. Kamegawa, M. Okada, Solid State Ionics, 154–155
(2002) 581-588
[108] I.N. Katsunori Akiyama, Masato Shida, Satoshi Ota, in, Mitsubishi Heavy
Industries, Ltd., (2009) US2007/0151481 A1.
[109] H. Anno, K. Matsubara, Y. Notohara, T. Sakakibara, H. Tashiro, Journal of
Applied Physics, 86 (1999) 3780-3786.
[110] S.-M. Choi, K.-H. Lee, C.-H. Lim, W.-S. Seo, Energy Conversion and
Management, 52 (2011) 335-339.
[111] R. Funahashi, M. Mikami, T. Mihara, S. Urata, N. Ando, Journal of Applied
Physics, 99 (2006) 066117.
[112] S. Urata, R. Funahashi, T. Mihara, A. Kosuga, S. Sodeoka, T. Tanaka, Int J
Appl Ceram Tec, 4 (2007) 535-540.
Page 48
Chapter III. Experimental Details
Chapter III. Experimental Details
3.1
POWDER PREPARATION
The starting materials for the preparation of oxide powders were La2O3
(99.99% purity, Sigma Aldrich), SrCO3 (99.9% purity, Sigma Aldrich), TiO2
(anatase) (99.6% purity, Alfa aesar) and Nb2O5 (99.9% purity, Alfa aesar). Different
powders according to their stoichiometric ratios were added to Nylon pots with ZrO2
balls as the grinding agent and ethanol as the milling medium. A mixture of 5 mm
and 10 mm diameter ZrO2 balls was used. The loaded pots were put in a planetary
ball mill (QM-3SP4, Nanjing University Instrument Plant, China) as shown in Figure
3.1(a) and rotated at a speed of 350 rpm for 2 hours to mix the powders. The ratio of
balls to powder was 10:1. After mixing, the powders were dried overnight at 80 oC in
a drying oven (Elite). After drying, the powders were sieved through a stainless steel
sieve with an aperture size of 250 µm.
In order to synthesise the required composition, the solid state reaction
method was employed. Mixed dried powders were put in to alumina crucible and
calcined at 1250-1300 oC for 4 hours using a chamber furnace (Carbolite HTF 1800)
as shown in Figure 3.1(b). After calcination, the synthesised powders were remilled
for 4 hours at 350 rpm to break the agglomerates and reduce the particle size. After
remilling, the powders were dried overnight at 80 oC and passed through a 250 µm
mesh sieve to control the particle size.
Page 50
Chapter III. Experimental Details
(a)
(b)
Figure 3.1: (a) QM-3SP4 Planetary ball mill machine (b) carbolite HTF 1800 Chamber furnace
For the synthesis of skutterudite powder, mechanical alloying technique was
used. Commercially available powders of Co (99.8% pure), Sb (99.5% pure), Te
(99.999% pure) and Yb (99.8% pure) were added to steel jars according to the
required stoichiometric ratios. This process was carried out in an Ar filled glove box
(as shown in Figure 3.2) to avoid oxidation of the powders. The jars were sealed
under a vacuum and loaded into the ball mill for 40 hours at 350 rpm. After mixing,
the steel jars were put inside the glove box to avoid oxidation of the powder.
Figure 3.2: Saffron scientific glove box used for powder handling under inert environment
Page 51
Chapter III. Experimental Details
3.2
Sintering by spark plasma sintering
Spark plasma sintering (SPS) is an efficient technique to sinter non
equilibrium, nanosized material by using high current pulsed DC and uniaxial
pressure. The main advantage of SPS is that it is a rapid sintering technique which
helps to retain the nanosize grains of the powders and lowering thermal conductivity
by orienting the grains in a preferred direction.
Figure 3.3: Spark Plasma Sintering furnace (FCT, Germany) with a schematic of its chamber
In this study, an SPS furnace (25/1 GCT, Germany) was used and shown in
Figure 3.3. It can achieve a heating rate of 600 oC/minute with a maximum
temperature of 2200 oC. The high pressure combined with rapid heating rate, helps to
make material with properties that cannot be achieved with conventional sintering. In
SPS, the grain growth is suppressed due to rapid sintering which helps to reduce
thermal conductivity. Apart from maintaining the grain size, SPS can also impart
defects in the structure which further helps to reduce thermal conductivity [1].
Page 52
Chapter III. Experimental Details
Calcined oxide powders were sintered using a SPS furnace. A set of graphite
dies and punches were used as a mould. Graphite dies and punches were
manufactured by Erodex (ISO 63). A graphite foil of 0.35 mm (supplied by SGL
Sigmaflex TH Foil) was used as a conducting barrier between graphite dies (and
punches) and the powder to avoid contamination and reaction. The powders were
cold pressed with a 0.5 tonne force. The cold pressed powders were sintered at 1350
o
C under 50 MPa of pressure for 5 minutes. The vacuum level inside the furnace was
~ 5 Pa.
For skutterudite, the powder was loaded into the graphite dies inside a glove
box and then sintered at 600 oC under Argon environment at 50 MPa pressure.
(a)
(b)
Figure 3.4: Schematic representation of: (a) SPS Die set; (b) photo of the graphite die during SPS
process
3.3
Post Sintering treatment
The sintered discs of oxide ceramics were annealed at 1000 oC in air for 6
hours in a chamber furnace (Carbolite HTF 1800) to remove any diffused carbon
during SPS processing. After air annealing, the ceramic samples were colourless and
Page 53
Chapter III. Experimental Details
electrically insulating. In order to enhance electrical conductivity, the ceramic
samples were annealed in an atmosphere of 10 % H2 + 90 % Ar. The samples were
annealed at 1250 oC for 4 hours in a thermal technology LLC furnace (Model
number 1100 2560 1/2). No post sintering treatment was done on skutterudites. The
samples were then ground to even out the surfaces using a Struers TegraPol-21
grinder/polisher. For electrical resistivity and Seebeck coefficient measurement,
ceramic samples were cut in to rectangular bars with dimensions of 3x3x15 mm
using an Accutom-5, Struers Cutting machine.
3.4
Characterization
3.4.1
Density measurement
The bulk densities of the sintered and air annealed and reduced ceramic discs
were measured by using the Archimedes principle and using the following relation
[2]
(3.1)
Where m1 is the mass of the sample in air, m2 is the mass of water soaked sample, m3
is the mass of sample inside water, ρ0 is the density of water and ρ is the density of
the sample. For nearly fully dense samples or the samples containing closed
porosity, m1 = m2 and the relation then become:
Page 54
Chapter III. Experimental Details
(3.2)
3.4.2
X-ray Diffraction and X-ray Photoelectron Spectroscopy
X-ray diffraction (XRD) is an analytical technique providing detailed
structural and compositional information of a material. It is a non-destructive
technique based on Bragg’s law. XRD patterns for the powders, sintered, air
annealed and hydrogen annealed ceramics were recorded using Siemens D5000 Xray diffractometer with Ni filtered Cu Kα radiation (λ = 1.54Å). Data were recorded
in the 2θ range of 100-700 with a step width of 0.03340 and a count time of 200
seconds per step.
All the diffraction patterns were collected by experimental officer Dr. Rory
M. Wilson. Powder samples were sprinkled on top of a zero background single
crystal silicon substrate for the analysis. The bulk ceramics were ground to a flat
surface and then mounted for recording their XRD pattern. Phase identification was
carried out by using PANalytical’s X’Pert HighScore version 2.1software.
In order to analyse the surface bond and chemical state of the elements, an
ESCALAB 250 X-ray Photoelectron Spectrometer (Thermo Corp.) was employed.
The equipment is set up at Aston University Birmingham. The equipment had a pass
energy of 20 eV and is equipped with monochromatized Al Kα (1486.5 eV) X-ray
source. Carbon was used as a reference material and XPSPEAK41 was used for
fitting XPS spectra.
Page 55
Chapter III. Experimental Details
3.4.3
Scanning and Transmission Electron Microscopy
The microstructure of the bulk ceramics were determined by Scanning
Electron Microscopy (SEM) (FEI, Inspect F). Samples for SEM were prepared by
grinding and polishing using silicon carbide papers up to grade 4000. After
polishing, all of the samples were thermally etched inside the chamber furnace to
reveal the grain boundaries. After thermal etching, the samples were cleaned with
acetone and gold coated to make them conductive.
Transmission Electron Microscope (TEM Jeol JEM 2010) was used to study
the detailed microstructure of the bulk samples. Samples for TEM were prepared by
grinding and polishing the bulk samples down to a thickness of ~ 30-50 µm. The
samples for polishing were mounted on to a transparent piece of glass with adhesive
‘crystal bond 590’. After polishing, the samples were immersed in acetone and
ultrasonicated for 10 minutes to dissolve crystal bond 590 and detach the sample
from glass substrate. The samples were then mounted onto nickel grid having an
aperture size of 1 mm with Araldite® adhesive. Before analysis, the samples were
left to dry overnight. After drying, the samples were put inside a Precision Ion
polishing System (PIPSTM) to reduce the thickness of the sample to electron
transparency. Image J analysis software was used to analyse the images and the
diffraction patterns obtained through TEM.
3.4.4
Electrical Characterisation
The samples were cut into 3x3x15 mm bars from the ceramic disc for
electrical resistivity and Seebeck coefficient measurements via four-point probe and
temperature differential methods. All the measurements were done in laboratoryPage 56
Chapter III. Experimental Details
made apparatus under vacuum. The equipment is shown in Figure 3.5. The
equipment was tested for calibration by comparing the test results with PbTe sample
supplied by the Laboratory of Thermoelectrics, Ben-Gurion University of the Negev,
Israel [3]. Measurements were carried out from room temperature to 500oC at an
interval of 100 oC. A heating rate of 10 oC/minute was employed to raise the
temperature to the test temperature. All the samples were rerun to check the
reproducibility of the measurements.
Figure 3.5: Lab built (in China) electrical resistivity and Seebeck coefficient measurement system
3.4.5
Thermal Characterisation
For the measurement of thermal conductivity the following relationship was
used,
Page 57
Chapter III. Experimental Details
κ= ρCpD
(3.3)
Where κ is thermal conductivity, ρ is the density, Cp is the specific heat and D is the
thermal diffusivity. Specific heat was measured by using a Netzsch STA (449 F3
Jupiter®). A sapphire disc (trade name white Sapphire No 12
⁄
boules) supplied by
Stetcher Thun GmbH with thickness of ~ 1 mm and diameter of 3 mm was used to
calibrate the instrument and the resulting data was within the error range of 7- 9 %.
The samples for the measurement was prepared by cutting a 3x3 mm square from the
bulk and polishing the bulk samples to a thickness of ~ 1 mm. All the measurements
were carried out in a N2 environment.
D is the thermal diffusivity which is ‘how fast a heat wave travels across a
certain area’; it has units of mm2/s. Thermal diffusivity was measured using a
netszch LFA 457 microflash as shown in Figure 3.6. The equipment was tested for
calibration with the standard samples provided by the manufacturer and the acquired
data was within the acceptable limit of the error (±3%). All the samples used had a
diameter of 20 millimetres with thickness of ~ 2.5-3 mm. The surfaces were ground
flat but left unpolished to minimise laser reflectance. A Kontakt-chemie Graphite 33
spray was used to coat both surfaces of the sample to maximise the absorption of the
laser. A laser beam diameter of 25.4 mm was used as there was no provision for 20
mm diameter laser beam. All the measurements were carried out under Ar
environment with a flow rate of 100 ml/min. Measurements were carried out from
room temperature to 800oC at 100oC intervals. A heating rate of 10oC/min was used
to raise the temperature within 10oC of the testing temperature, after which, a heating
rate of 1oC/minute was used to achieve the testing temperature. The maximum
Page 58
Chapter III. Experimental Details
temperature fluctuations were set to ±1oC/minute In order to enhance accuracy and
reduce the scatter, three measurements per temperature point were recorded.
Figure 3.6: Netzsch laser flash LFA 457
Page 59
Chapter III. Experimental Details
REFERENCES
[1] S. Ballikaya, H. Chi, J.R. Salvador, C. Uher, Journal of Materials Chemistry A,
1 (2013) 12478-12484.
[2] R.M. German, in: Sintering Theory and Practice, John Wiley & Sons, Inc., New
York, 1996.
[3] Y. Gelbstein1, J. Davidow1, S. N. Girard, D. Young Chung and M. Kanatzidis,
Journal of Advanced Energy Materials, 3 (2013) 815-820.
Page 60
Chapter IV. Characterization of La4Ti4O14±δ
Chapter IV. Characterization of La4Ti4O14±δ
4.1
INTRODUCTION
The aim of this research was to investigate the effect of acceptordonor substitution and oxygen stoichiometry on the thermoelectric properties
(especially thermal conductivity) of the PLS La4Ti4O14. PLS ceramics tend to have
very low thermal conductivity (typically ~ 1-2 W/m.K) [1, 2]. Since polycrystalline
La4Ti4O14 is an electrical insulator with high electrical resistivity (~ 1015 Ωcm) [3],
reduction by hydrogen may cause a partial valence change of Ti from +4 to +3,
which produces more conduction electrons. As the temperature increases, more and
more electrons reach the conduction band thus decreasing electrical resistivity. Also,
by substituting La or Ti in the crystal structure with suitable cations, it creates
compositional non-stoichiometry and mass contrast which affects the thermal
conductivity [4, 5]. There is no reported literature on the thermal conductivity of
La4Ti4O14 ceramics. In the current research Sr, Ta and Nb substituted La4Ti4O14 were
synthesized and their microstructures were characterized in relation to compositional
non-stoichiometry.
Page 61
Chapter IV. Characterization of La4Ti4O14±δ
4.2
EXPERIMENTAL DETAILS
La4-xSrxTi4O14-
±δ,
La4Ti4-xTaxO14+
±δ
and La4Ti4-xNbxO14+
±δ
(where
x=0, 0.2, 0.4, 0.6 and 0.8), were prepared by Solid State Reaction. The starting
materials were La2O3 (99.99% purity, Sigma Aldrich), SrCO3 (99.9% purity, Sigma
Aldrich) Nb2O5 (99.9% Alfa Aesar) and TiO2 (anatase) (99.6% purity, Alfa Aesar).
These powders were mixed in stoichiometric ratios and ball milled using ethanol as
the milling medium in a planetary ball mill. Mixed powders were calcined at 1300oC
for 4 hours. After calcination, the powders were remilled for 4 hours to break the
agglomerates and reduce the particle size.
Calcined powders were sintered using a SPS furnace at a heating rate of
100oC/min. The powders were cold pressed into a 20 millimetre graphite die and
sintered at 1350oC under 50 MPa of pressure for 5 minutes. Figure 4.1 shows typical
SPS processing parameters for the pure and doped La4Ti4O14.
Page 62
Chapter IV. Characterization of La4Ti4O14±δ
a)
b)
Figure 4.1: Typical processing parameters during sintering of: (a) La4Ti4O14±δ; (b)
La3.2Sr0.8Ti4O13.6±δ
Page 63
Chapter IV. Characterization of La4Ti4O14±δ
Figure 4.2 shows photos of ceramic disc of sintered La4Ti4O14. The disc just
after SPS was black (Figure 4.2(a)) because of carbon contamination from the
graphite dies/ punches and the low level of vacuum (~ 5 Pa). The sintered discs were
air annealed at 1000oC for 6 hours to remove any diffused carbon during SPS
processing. After air annealing the ceramic disc were colourless (Figure 4.2(b)). The
air annealed discs were reduced at 1200oC for 4 hours in a mixture of 10 % H2 90%
Ar to increase the electrical conductivity and the disc again turned black (Figure
4.2(c)).
a)
b)
c)
Figure 4.2: Photos of ceramic disc: (a) after sintering; (b) after air annealing; (c) after reduction
Page 64
Chapter IV. Characterization of La4Ti4O14±δ
RESULTS AND DISCUSSIONS
4.3
La4-xSrxTi4O14- ±δ
Table 4.1 shows the theoretical and measured densities of the
La4-xSrxTi4O14-
±δ
compositions. The bulk density was measured by the Archimedes
method and theoretical density was calculated by
(4.1)
Where M is the average molar mass, Z is the number of atoms per unit cell, ω
is the unit cell volume and NA is the Avogadro’s number. All the sintered samples
were more than 98% dense and the theoretical density decreased with increasing Sr
substitution.
Table 4.1: Theoretical and relative densities of La4-xSrxTi4O14-
±δ ceramics
Theoretical Density Relative Density (%)
(gcm-3)
La4Ti4O14±δ
5.78
99.3
La3.8Sr0.2Ti4O13.9±δ 5.71
99.4
La3.6Sr0.4Ti4O13.8±δ 5.64
99.6
La3.4Sr0.6Ti4O13.7±δ 5.57
99.8
La3.2Sr0.8Ti4O13.6±δ 5.50
99.8
Page 65
Chapter IV. Characterization of La4Ti4O14±δ
Figure 4.3(a) shows the X-ray diffraction patterns for La4-xSrxTi4O14-
±δ
(x=0,
0.2, 0.4, 0.6 and 0.8) powder. The peaks matched with La4Ti4O14 (LaTiO3.5), n=4
(PDF card # 28-0517). The peaks were shifted towards lower angles indicating that
the lattice parameter increased as a result of substitution of La3+ (ionic radius 1.17Å)
by Sr2+ (ionic radius 1.32Å). All the peaks were sharp giving an indication of a large
particle size of the powder according to the Scherrer formula [6]. A secondary phase
was also generated which cannot be indexed and marked as *.
After sintering the unidentified secondary phase disappeared and a new phase
was produced, the 5 layer perovskite compound La5Ti5O17, (PDF # 00-048-0480)
which is indicated with the symbol ♦ in Figure 4.3(b). The amount of La5Ti5O17
increased with increasing Sr content. La5Ti5O17 (LaTiO3.4) has a monoclinic
structure with the Space group Pc. After air annealing, the amount of La5Ti5O17
decreased slightly as shown in Figure 4.3(c). This is due to the fact that the
La5Ti5O17 is stable in low partial pressure of oxygen as it has a smaller oxygen ratio
with the A and B site elements compared to La4Ti4O14 (A:B:O is 1:1:3.5 for
La4Ti4O14 and 1:1:3.4 for La5Ti5O17) [7]. This effect is reversed slightly during
reduction as seen in Figure 4.3(d).
Page 66
Chapter IV. Characterization of La4Ti4O14±δ
a)
La3.2Sr0.8Ti4O13.6±δ
La3.8Sr0.2Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La4Ti4O14±δ
La4Ti4O14
La5Ti5O17
b)
La3.2Sr0.8Ti4O13.6±δ
La3.8Sr0.2Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La4Ti4O14±δ
La4Ti4O14
La5Ti5O17
Page 67
Chapter IV. Characterization of La4Ti4O14±δ
c)
La3.2Sr0.8Ti4O13.6±δ
La3.4Sr0.6Ti4O13.9±δ
La3.6Sr0.4Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La4Ti4O14±δ
La4Ti4O14
La5Ti5O17
d)
La3.2Sr0.8Ti4O13.6±δ
La3.4Sr0.6Ti4O13.9±δ
La3.6Sr0.4Ti4O13.9±δ
La3.8Sr0.2Ti4O13.9±δ
La4Ti4O14±δ
La4Ti4O14
La5Ti5O17
Figure 4.3: XRD patterns of the La4-xSrxTi4O14- ±δ: (a) powders; (b) sintered; (c) air annealed; (d)
reduced samples
Page 68
Chapter IV. Characterization of La4Ti4O14±δ
The amount of La5Ti5O17 was calculated using the Normalized Relative
Intensity Ratio (RIR) method proposed by Chung [8] and is presented in Table 4.2.
Table 4.2: amount of 5 layer PLS compound in La4-xSrxTi4O14Sinter (mass %)
±δ ceramics
Air Annealed (mass %)
Reduced (mass %)
La3.8Sr0.2Ti4O13.9±δ -
-
-
La3.6Sr0.4Ti4O13.8±δ 5.4
2.1
2.6
La3.4Sr0.6Ti4O13.7±δ 6.8
2.2
3.4
La3.2Sr0.8Ti4O13.6±δ 8.4
3.0
7.8
Figure 4.4 shows typical scanning electron microscope (SEM) images of
calcined La4-xSrxTi4O14-
±δ
powder before and after ball milling. After calcination,
some of the powder was agglomerated with a particle size of ~ 2 µm. The ball
milling process broke the agglomerates and reduced the average particle size to ~ 0.7
µm and the particles became more homogeneous.
a)
b)
Figure 4.4: Typical SEM images of La3.2Sr0.8Ti4O13.9±δ powder after calcination: (a) before ball
milling; (b) after ball milling
Page 69
Chapter IV. Characterization of La4Ti4O14±δ
SEM images of polished and etched La4-xSrxTi4O14-
±δ
ceramics (x=0.2-0.8
sintered at 1350°C are shown in Figure.4.5. The plate like nature of the La4Ti4O14
grains decreased with the addition of Sr which is in consistent with a previous report
in which the plate-like nature of La4Ti4O14 was suppressed by the addition of Ce [3].
There were still some grains that showed plate-like behaviour. The grain size was
typically in the range of 0.3-0.7 µm with a few larger grains.
a)
b)
c)
d)
Figure 4.5: SEM images of LST ceramics after polishing and etching: (a) La3.8Sr0.2Ti4O13.9±δ; (b)
La3.6Sr0.4Ti4O13.8±δ; (c) La3.6Sr0.6Ti4O13.7±δ; (d) La3.2Sr0.8Ti4O13.6±δ
Page 70
Chapter IV. Characterization of La4Ti4O14±δ
Figure 4.6(a) shows a typical bright field TEM image of the La4Ti4O14
ceramics after air annealing. The typical grain size of the ceramic was ~ 0.7 µm and
most of the grains were defect free except for a few planar defects. Figure 4.6(b)
shows a high resolution image of (100) lattice planes. The inter-planar spacing
corresponds to the 4 layer La4Ti4O14 (~ 12.63Å). The corresponding diffraction
pattern is shown in Figure 4.6(c). Figures 4.6(d-e) show bright field TEM images of
La3.2Sr0.8Ti4O13.6±δ air annealed ceramic, which has a large number of nanoscale
intergrowths. These intergrowths were distributed homogeneously throughout the
ceramic. Some of the intergrowths terminated or moved to a different plane of the
La4Ti4O14 as shown in the Figure 4.6(f). When two intergrowths terminated, they
disturbed the neighbouring planes and created a ‘bulging’ of the planes in between as
clearly seen in the Figure 4.6(f). The inter-planar spacing also changed when there
was an intergrowth. The spacing difference before and after the intergrowth
corresponds to one octahedral layer in the PLS compounds, which means the
structure is changing from a La4Ti4O14 (4 layer, d spacing = 12.63 Å) to a La5Ti5O17
(5 layer, d spacing of 15.7 Å) perovskite compound as a result of intergrowth of the
5 layer PLS compound as shown in Figure 4.6(g-h).
Page 71
Chapter IV. Characterization of La4Ti4O14±δ
b)
a)
c)
d)
e)
f)
g)
h)
Figure 4.6: TEM images of air annealed: (a) La4Ti4O14±δ; (b) high magnification image of
La4Ti4O14±δ;(c) diffraction pattern of the air annealed La4Ti4O14; (d) La3.2Sr0.8Ti4O13.6±δ showing the
large defect (intergrowth) number; (e) La3.2Sr0.8Ti4O13.6±δ; (f) La3.2Sr0.8Ti4O13.6±δ showing termination
of intergrowths; (g-h) La3.2Sr0.8Ti4O13.6±δ showing two regions of intergrowths
Page 72
Chapter IV. Characterization of La4Ti4O14±δ
Figure 4.7(a) shows a bright field image of reduced La4Ti4O14 based ceramic.
After reduction the density of La4Ti4O14 decreased slightly (from 5.74 g cm-3to5.62 g
cm-3) and few planar defects appeared. Figure 4.7(b) represents high resolution
image of a typical planar defect in reduced La4Ti4O14±δ ceramic. After reduction the
density of La3.2Sr0.8Ti4O13.6±δ decreased slightly (from 5.53 g cm-3to 5.47 g cm-3).
The structure still has a high density of intergrowths as shown in Figure 4.7(c-d). It
is inconclusive to comment on difference in intergrowth number between air
annealed and reduced ceramics from the TEM images, but the XRD clearly indicated
an increased amount of La5Ti5O17 intergrowths in reduced La3.2Sr0.8Ti4O13.6±δ. Figure
4.7(e) shows a magnified region of intergrowths and their termination. The
corresponding diffraction pattern is also shown in Figure 4.7 (f). The structure still
shows bulging of planes as observed in Figure 4.6 (e and f).
a)
b)
c)
d)
Page 73
Chapter IV. Characterization of La4Ti4O14±δ
e)
f)
Figure 4.7: TEM images of reduced: (a) La4Ti4O14±δ; (b) high magnification image of La4Ti4O14±δ;
(c-d) La3.2Sr0.8Ti4O13.6±δ showing the large defect (intergrowth) number;(e) a high magnification
image of a typical region in La3.2Sr0.8Ti4O13.6±δ showing intergrowths; (f) corresponding diffraction
pattern of La3.2Sr0.8Ti4O13.6±δ
Figure 4.8(a–d) shows the XPS spectra of surface electrons for different
elements in La4Ti4O14 based ceramics after ion beam etching. The ion beam etching
was done for 5 min. Since the C reference used by different labs varies, the absolute
value of each element varies in a range. Figure 4.8(a) shows the XPS spectra of La
3d electrons. The binding energy (BE) values for La 3d showed a clear doublet for
both components (La3d5/2 and La3d3/2) which is a characteristic of La containing
oxide compounds [9]. The binding energy difference between the splitting of La
3d5/2 and La 3d3/2 was ~ 4.6eV and ~ 4.4 eV, respectively, for all of the ceramics.
Figure 4.8(b) shows the XPS spectra of a typical Ti 2p doublet. For air annealed
La4Ti4O14, the peak position marked as A (459.17 eV) and the FWHM match the
Ti4+ state [3]. The Ti 2p peak is slightly asymmetric and it gives rise to another peak
at 457.67 eV which is marked as B. It is inconclusive to say whether this peak was
caused by the ion beam etching damage and (or) because of the presence of Ti3+ state
[10]. After reduction, one additional peak can be seen at 453.9 eV which is marked
as C. This peak can be attributed to the Ti2+ state as reported in literature [11, 12].
The intensity of the peak marked as Bˈ also increased in the reduced ceramic. Since
Page 74
Chapter IV. Characterization of La4Ti4O14±δ
both the samples were subjected to the same amount of etching; it is conclusive to
determine the presence of Ti3+ state also.
The corresponding O 1s core level spectra are shown in the Figure 4.8(c). All
the spectra show same symmetry towards the lower binding energy side which can
be attributed to the bonded oxygen in the structure as reported in literature [13]. The
higher binding energy component indicates chemisorbed oxygen or absorbed
hydroxyl ions [9]. In the case of the air annealed La4Ti4O14±δ, the peak position for
the oxygen was 530.7 eV which changed to 529.6 eV after hydrogen reduction. This
decrease in binding energy can be related to the loss of oxygen or the oxygen being
compromised during the formation of nanoscale intergrowths.
Figure 4.8(d) gives the Sr 3d core level spectrum. The peak shape and
position matched well with the reported value [13]. After Sr addition, there was no
noticeable difference in the peak size and shape for La3d electrons, the binding
energy difference between La 3d5/2 and La 3d3/2 doublets in air annealed
La3.2Sr0.8Ti4O14±δ was 4.44 eV. For Ti 2p electrons, the peak marked as Bˈˈincreased
which indicates the presence of Ti3+ ions. Since the TEM and XRD results showed
the presence of localized 5 layer structure, the decrease in binding energy can be
related to an increase in separation of the oxygen interlayer. For O1s, the peak
position was at 530.4 eV which is on the lower side as compared to air annealed
La4Ti4O14±δ (529.6 eV). Since 5 layer PLS compound has less oxygen content than 4
layer PLS compound and the amount of 5 layer intergrowths increase in the air
annealed La3.2Sr0.8Ti4O13.6±δ ceramic as compared to La4Ti4O14±δ. This indicates that
the oxygen has been compromised during the formation of an intergrowth.
Page 75
Chapter IV. Characterization of La4Ti4O14±δ
a)
La3.2Sr0.8Ti4O13.6±δ
b)
La3.2Sr0.8Ti4O13.6±δ air annealed
Page 76
Chapter IV. Characterization of La4Ti4O14±δ
c)
La4Ti4O14±δ
air annealed
La4Ti4O14±δ
reduced
La3.2Sr0.8Ti4O13.6±δ
air annealed
d)
La3.2Sr0.8Ti4O13.6±δ
air annealed
Figure 4.8: XPS spectrum of La4-xSrxTi4O13.6±δ: (a) La 3d; (b) Ti 2p; (c) O 1s; (d) Sr 3d
Page 77
Chapter IV. Characterization of La4Ti4O14±δ
In order to improve the quantitative analysis of the data, the binding energy
difference (ΔBE) method was used which helps to reduce scatter in the data [14].
The ΔBE values for the pure and Sr substituted La4Ti4O14±δ are given in Table 4.3.
The ΔBE (O-Ti) for air annealed and hydrogen reduced La4Ti4O14±δ samples was
calculated to be 71.5 eV and 71.6 eV respectively. These values are in good
agreement with the published literature [11]. The ΔBE (O-Ti) does not change after
the addition of Sr. The ΔBE (La-O) for air annealed La4Ti4O14 was found to be
304.7 eV which decreased to 304.2 eV after reduction. This decrease in binding
energy was due to the decrease in bonding as a result of oxygen removal after
reduction. After the Sr substitution, the ΔBE (La-O) was 304.4 eV which is lower
than the pure air annealed La4Ti4O14 but higher than the pure reduced La4Ti4O14±δ.
These results are in accordance with the TEM results which suggest the presence of
localised planar defects/intergrowths in the reduced and Sr substituted La4Ti4O14
which causes the weakening of the La-O bonds. ΔBE (Sr-O) was calculated to be
397 eV and it agrees with the reported value [13].
Table 4.3: XPS data of La4-xSrxTi4O13.6±δ ceramics
Material
Peak Position (eV)
Binding Energy Difference
(∆eV)
La3d
La3d3/2
Ti2P
O1s
La-O
Ti-O
Sr-O
La3d5/2
La4Ti4O14±δ air annealed 852.2
835.4
459.2
530.7
304.7
71.5
-
La4Ti4O14±δ reduced
850.6
833.8
458
529.6
304.2
71.6
-
air 851.6
834.8
458.8
530.4
304.4
71.6
397.0
La3.2Sr0.8Ti4O13.6±δ
annealed
Page 78
Chapter IV. Characterization of La4Ti4O14±δ
4.4
La4Ti4-xTaxO14+ ±δ
Table 4.4 presents the theoretical and relative densities of pure and
La4Ti4-xTaxO14+
±δ
ceramics after SPS. The theoretical density of all the ceramics
increased after substitution due to higher atomic weight of Ta (~ 180.9 amu). All the
ceramics were more than 98% dense.
Table 4.4: Theoretical and relative densities of La4Ti4-xTaxO14+
±δ
ceramics
Theoretical Density Relative Density (%)
(g cm-3)
La4Ti4O14±δ
5.78
99.3
La4Ti3.9Ta0.1O14.05±δ 5.87
99.4
La4Ti3.8Ta0.2O14.1±δ
5.95
99.2
La4Ti3.6Ta0.4O14.2±δ
6.12
97.2
La4Ti3.4Ta0.6O14.3±δ
6.29
96.5
Figure 4.9 shows the X ray diffraction patterns for La4Ti4-xTaxO14+
±δ
(x=0,
0.1, 0.2, 0.4 and 0.6) in powder, sintered, air annealed and reduced conditions. No
secondary phase was observed in the powders till x = 0.4 and the peaks matched
with La4Ti4O14, n=4 (PDF card # 28-0517). After x=0.4, a secondary phase was
produced as a result of Ta doping. This new generated phase was indexed as
La3Ti2TaO11 (PDF card # 00-054-0632) which is a n =3 PLS structure. All the peaks
in the XRD pattern were sharp giving an indication of a large particle size (> 0.5μm).
Page 79
Chapter IV. Characterization of La4Ti4O14±δ
a)
La4Ti3.4Ta0.6O14.3±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti4O14±δ
La4Ti4O14
La3Ti2TaO11
b)
La4Ti3.4Ta0.6O14.3±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti4O14±δ
La4Ti4O14
La3Ti2TaO11
Page 80
Chapter IV. Characterization of La4Ti4O14±δ
La
La4Ti
Ta
Ta0.60.6OO14.3±δ
4Ti
3.4
3.4
14.3±δ
c)
La
La44Ti
Ti3.6
Ta0.4
O14.2±δ
3.6Ta
0.4O
14.2±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti4O14±δ
La4Ti4O14
La3Ti2TaO11
2 Theta (deg.)
La4Ti3.4Ta0.6O14.3±δ
d)
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti4O14±δ
La4Ti4O14
La3Ti2TaO11
2 Theta (deg.)
Figure 4.9: XRD patterns of the La4Ti4-xTaxO14+ ±x: (a) powders; (b) sintered; (c) air annealed; (d)
reduced samples
Page 81
Chapter IV. Characterization of La4Ti4O14±δ
For La4Ti3.4Ta0.6O14.3±δ powder, the amount of La3Ti2TaO11 was estimated to be ~
3.6 mass % which increased to ~ 4.6 mass % for sintered and air annealed ceramics.
The amount of generated secondary phase was reduced to ~ 1.7 mass % after
reduction. The reduced ceramics have less oxygen content and as a result the amount
of second phase decreased. The decreased intensity of peak (♦) in reduced
La4Ti3.4Ta0.6O14.3±δ ceramic clearly indicated the reduction in the amount of
secondary phase. The peak intensity for the secondary phase was weak in the other
compositions (x=0.1, 0.2 and 0.4) which did not allow the amount of secondary
phase to be calculated.
Figure 4.10(a-b) shows a bright field TEM image of La4Ti3.6Ta0.4O14.2±δ air
annealed ceramic with some planar defects. These defects were distributed
homogeneously throughout the ceramic as shown in Figure 4.10(c). From XRD data,
the formation of 3 layered PLS compound was confirmed. So, the regions of planar
defects are the regions of 3 layered phase embedded inside 4 layered La4Ti4O14. The
structure of a 3 layer PLS compounds exist in two types, Type I and Type II. In Type
I structure, the unit cell contains regular stacking of 3 layers of BO6 octahedra, while
in Type II, the unit cell contains alternate stacking of 4 and 2 layers of BO6
octahedra. La3Tia2TaO11 exists in the Type II structure as reported by Titov et al.
[15]. The intergrowths in La4Ti4-xTaxO14+
La4-xSrxTi4O14-
±δ.
±δ
are different from the ones in
as shown in Figure 4.10(d). The intergrowths are shorter and they
are within the 4 layer PLS compound. Figure 4.10(e) shows a magnified image of
the region with intergrowths and the corresponding diffraction pattern is shown in
Figure 4.11(f).
After reduction, the density of La4Ti3.6Ta0.4O14.2±δ decreased slightly (from
6.07 g cm-3to 5.98 g cm-3) and the density of intergrowths also decreased as observed
Page 82
Chapter IV. Characterization of La4Ti4O14±δ
in the X-ray diffraction data (mass % decreased from 4.6 to 1.7). It is inconclusive to
quantitatively analyse the defect density difference between air annealed and reduced
La4Ti3.6Ta0.4O14.2±δ from TEM.
a)
b)
c)
d)
e)
f)
Figure 4.10: TEM images of: (a) air annealed La4Ti3.6Ta0.4O14.2±δ; (b) high magnification image
of air annealed La4Ti3.6Ta0.4O14.2±δ; (c) reduced La4Ti3.6Ta0.4O14.2±δ; (d) high magnification image of
reduced d La4Ti3.6Ta0.4O14.2±δ (e) high magnification image of reduced d La4Ti3.6Ta0.4O14.2±δ with d
spacing; (f) diffraction pattern of high magnification image of reduced d La4Ti3.6Ta0.4O14.2±δ
Page 83
Chapter IV. Characterization of La4Ti4O14±δ
The nanoscale intergrowths observed in the high magnification TEM images
of air annealed La4Ti3.6Ta0.4O14.2±δ match exactly with the XRD results of air
annealed La4Ti3.6Ta0.4O14.2±δ ceramic, where the peak width of all the peaks has
increased. This peak broadening is attributed to the localized intergrowths of 3 layer
PLS phase within the 4 layer structure. The broadening of the peaks followed a
systematic trend increasing with increasing amount of Ta in La4Ti4O14 as given in
Table 4.5. After reduction, the peak width decreases and also the amount of
generated secondary phase.
Table 4.5: FWHM of the La4Ti4-xTaxO14+
Powder
Sinter
Annealed
±δ
Reduced
ceramics
Powder
Sinter
Annealed
Reduced
x
Peak (-212) /2θ0
0
0.14
0.15
0.16
0.14
0.14
0.16
0.17
0.18
0.1
0.12
0.14
0.16
0.11
0.14
0.14
0.18
0.11
0.2
0.14
0.13
0.20
0.13
014
0.14
0.19
0.13
0.4
0.13
0.12
0.23
0.13
0.14
0.13
0.20
0.13
0.6
0.14
0.15
0.31
0.14
0.14
0.14
0.26
0.15
(400) /2θ0
Figure 4.11 compares the XPS spectra of pure and Ta substituted La4Ti4O14.
The explanation of La4Ti4O14 air annealed and reduced data has been discussed
previously (section 4.3). After Ta addition, the binding energy for La 3d doublet and
the Ti 2p increased compared to those for La4Ti4O14 and are given in Table 4.6. The
B peak which was observed for Ti 2p in pure La4Ti4O14±δ disappeared after Ta
substitution. This indicates that Ti is present in Ti4+ form in Ta substituted
La4Ti4O14. The binding energy for O 1s is similar to the one in pure La4Ti4O14±δ.
Page 84
Chapter IV. Characterization of La4Ti4O14±δ
a)
La
4Ti3.6Ta0.4O14.2±δ
La
4Ti3.6Ta4O14.2±δ
air annealed
b)
La4Ti3.6Ta0.4O14.2±δ
air annealed
Page 85
Chapter IV. Characterization of La4Ti4O14±δ
c)
La4Ti3.6Ta0.4O14.2±δ
air annealed
d)
La4Ti3.6Ta0.4O14.2±δ
air annelaed
Figure 4.11: XPS spectrum of La4Ti4-xTaxO14+ ±δ: (a) La 3d; (b) Ti 2p; (c) O1s; (d) Ta 4f
Page 86
Chapter IV. Characterization of La4Ti4O14±δ
Table 4.6 compares the peak position and ∆BE values of all the elements of
La4Ti3.6Ta0.4O14.2±δ ceramic. The ∆BE (La-O) and ∆BE (O-Ti) increases after Ta
substitution in La4Ti4O14. Since both La and Ti exist in two bonding states, i.e. in
La4Ti4O14 and La4Ti3.6Ta0.4O14.2±δ, the increase in ∆BE indicates an increase in the
bonding between La/Ti and oxygen.
Table 4.6: XPS data of La4Ti3.6Ta0.4O14.2±δ ceramics
Material
Peak Position (eV)
Binding
Energy
Difference
(∆eV)
La3d
Ti2P
O1s
La-O
Ti-O
Ta-O
La3d3/2
La3d5/2
852.2
835.4
459.2
530.7
304.7
71.5
-
La4Ti4O14±δ reduced
850.6
833.8
458
529.6
304.2
71.6
-
La4Ti3.6Ta0.4O14.2±δ
853.0
836.2
458.4
530.6
305.6
72.2
504.4
La4Ti4O14±δ air
annealed
air annealed
Page 87
Chapter IV. Characterization of La4Ti4O14±δ
4.5
La4Ti4-xNbxO14+ ±δ
Table 4.7 represents the theoretical and relative densities of pure and Nb
substituted La4Ti4O14 ceramics after SPS. The theoretical density of all the ceramics
increased after substitution due to higher atomic weight of Nb (~ 92.9 amu). All the
ceramics were more than 98% dense.
Table 4.7: Theoretical and relative densities of La4Ti4-xNbxO14+
±δ ceramics
Theoretical Density Relative Density (%)
(g cm-3)
La4Ti4O14±δ
5.78
99.3
La4Ti3.8Nb0.2O14.1±δ 5.85
99.5
La4Ti3.6Nb0.4O14.2±δ 5.91
99.3
La4Ti3.4Nb0.6O14.3±δ 5.97
98.4
X ray diffraction patterns for La4Ti4-xNbxO14 (x=0, 0.2, 0.4 and 0.6) in
powder, sintered, air annealed and reduced conditions are shown in Figure 4.12. All
the peaks matched with La4Ti4O14, n=4 (PDF card # 28-0517) until x= 0.2. After
x=0.2, a second phase was generated which was indexed as LaNbO4 (PDF card # 01081-1973) which is a n = 2 PLS phase. The second phase was produced as a result of
the reaction between La and Nb during calcination. For x= 0.4, a shoulder at 28.8o (2
theta) gives an indication of the second phase. This shoulder increases with
increasing Nb content. The amount of LaNbO4 (by mass) was calculated by the RIR
method and was found to be ~ 3 mass% and ~ 0.9 mass % for La4Ti3.4Nb0.6O14.3±δ
powder and sintered La4Ti3.6Nb0.4O14.3±δ respectively.
Page 88
Chapter IV. Characterization of La4Ti4O14±δ
a)
La4Ti3.4Nb0.6O14.3±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti4O14±δ
La4Ti4O14
LaNbO4
b)
La4Ti3.4Nb0.6O14.3±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti4O14±δ
La4Ti4O14
LaNbO4
Page 89
Chapter IV. Characterization of La4Ti4O14±δ
c)
La4Ti3.4Nb0.6O14.3±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti4O14±δ
La4Ti4O14
LaNbO4
d)
La4Ti3.4Nb0.6O14.3±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti4O14±δ
La4Ti4O14
LaNbO4
Figure 4.12: XRD patterns of the La4Ti4-xNbxO14.3±δ: (a) powders; (b) sintered; (c) air annealed;
(d) reduced samples
Page 90
Chapter IV. Characterization of La4Ti4O14±δ
Bright field TEM images of La4Ti3.4Nb0.6O14.3±δ air annealed ceramic are
given in Figure 4.13 (a-b). As observed in section 4.4, Ta substitution in La4Ti4O14
generates a 3 layered PLS compound, but Nb substitution did not form a n=3 PLS
compound but formed a 2 layer PLS compound instead which was LaNbO4 in this
case. From the XRD data, formation of the n=2 layer PLS compound LaNbO4 was
confirmed and it can be clearly seen as a single phase grain in the Figure 4.13(b).
The microstructure of the LaNbO4 matches with that reported by Lee et al. [16]. The
d spacing (~ 2.8 Å) matches with the d spacing of the LaNbO4 phase as shown in
Figure 4.13(c). The corresponding diffraction pattern is shown in the Figure 4.13(d).
In some areas, LaNbO4 was uniformly distributed while in others, it was found in
clusters. The same is true for La4Ti4O14 grains as shown in Figure 4.13(b and d).
Figure 4.13(f) shows a high magnification image of a typical LaNbO4 grain. Some
planar defects were also seen in the grains of LaNbO4. The origin of these defects
cannot be determined due to the involvement of many variables which include SPS
processing or the substitution of Ti inside LaNbO4. After reduction, the density of
La4Ti3.4Nb0.6O14±δ decreased slightly (from 5.8 g cm-3to 5.76 g cm-3) and the amount
of LaNbO4 stayed nearly the same as observed in X-ray diffraction.
Page 91
Chapter IV. Characterization of La4Ti4O14±δ
a)
b)
c)
d)
e)
f)
Figure 4.13: TEM images of: (a) air annealed La4Ti3.4Nb0.6O14.3±δ; (b) air annealed
La4Ti3.4Nb0.6O14.3±δ showing region of LaNbO4; (c) typical grain of reduced La4Ti3.4Nb0.6O14.3±δ
ceramic showing LaNbO4 grain; (d) corresponding diffraction pattern of LaNbO4 grain; (e) a typical
region showing cluster of La4Ti4O14 grains in reduced ceramic; (f) a typical LaNbO4 grain showing
planar defects in reduced ceramic
Page 92
Chapter IV. Characterization of La4Ti4O14±δ
4.6
CONCLUSION
Sr, Ta and Nb substituted La4Ti4O14 ceramics were synthesized by solid state
reaction and densified by spark plasma sintering. The effect of substitution and
oxidation-reduction on phase contrast and microstructure was studied. Substitution
of elements in the structure produced disorder in the structure and additional heat
treatment did not produce well-ordered structures. Acceptor (Sr) substitution
produced nanoscale intergrowths of 5 layer PLS compound while donor (Ta)
substitution produced 3 layer PLS compound. These intergrowths were different in
size and appearance in acceptor and donor substituted La4Ti4O14. These nanoscale
intergrowths were caused by the localized disturbance in the octahedral layers.
Evidence of 5 or 3 layer PLS compounds was confirmed by XRD and transmission
electron microscopy. Nanoscale intergrowths number increased in the case of
acceptor substitution after reduction while decreased in the case of donor
substitution. Nb substitution in La4Ti4O14 did not produce 3 layer PLS compound but
produced 2 layer PLS compound LaNbO4 which was observed as a separate phase
inside the microstructure.
Page 93
Chapter IV. Characterization of La4Ti4O14±δ
REFERENCES
[1] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
[2] A. Sakai, T. Kanno, K. Takahashi, Y. Yamada, H. Adachi, Journal of Applied
Physics, 108 (2010) 103706.
[3] Z.P. Gao, H.X. Yan, H.P. Ning, R. Wilson, X.Y. Wei, B. Shi, H. Ye, M.J. Reece,
Journal of the European Ceramic Society, 33 (2013) 1001-1008.
[4] D.R. Clarke, Surface and Coating Technology, 163 (2003) 67-74.
[5] A. Rauf, Q. Yu, L. Jin, C. Zhou, Scripta Materialia, 66 (2012) 109-112
[6] P. Scherrer, Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen,
Mathematisch-Physikalische Klasse, 2 (1918) 98.
[7] T. Williams, H. Schmalle, A. Reller, F. Lichtenberg, D. Widmer, G. Bednorz,
Journal of Solid State Chemistry, 93 (1991) 534-548.
[8] F.H. Chung, J Appl Crystallogr, 8 (1975) 17-19.
[9] V.V. Atuchin, T.A. Gavrilova, J.C. Grivel, V.G. Kesler, Journal of Physics D:
Applied Physics, 42 (2009) 035305.
[10] S. Hashimoto, A. Tanaka, Surface and Interface Analysis, 34 (2002) 262-265.
[11] L. Bugyi, A. Berkó, L. Óvári, A.M. Kiss, J. Kiss, Surface Science, 602 (2008)
1650-1658
[12] W.S. Oh, C. Xu, D.Y. Kim, D.W. Goodman, Journal of Vacuum Science &
Technology A, 15 (1997) 1710-1716
[13] Z. Gao, H. Ning, C. Chen, R. Wilson, B. Shi, H. Ye, H. Yan, M.J. Reece, J.L.
Jones, Journal of the American Ceramic Society, 96 (2013) 1163-1170.
Page 94
Chapter IV. Characterization of La4Ti4O14±δ
[14] V.V. Atuchin, T.A. Gavrilova, J.C. Grivel, V.G. Kesler, Surface Science, 602
(2008) 3095-3099.
[15] Y.A. Titov, A.M. Sych, V.Y. Markiv, N.M. Belyavina, A.A. Kapshuk, V.P.
Yaschuk, Journal of Alloys and Compounds, 316 (2001) 309-315.
[16] H.-W. Lee, J.-H. Park, S. Nahm, D.-W. Kim, J.-G. Park, Materials Research
Bulletin, 45 (2010) 21-24.
Page 95
Chapter V. Characterization of Sr4Nb4O14±δ
Chapter V. Characterization of Sr4Nb4O14±δ
5.1
INTRODUCTION
Sr4Nb4O14 is a PLS ceramic known for its ferroelectric and
piezoelectric properties [1]. Pure Sr4Nb4O14 is a wide band gap semiconductor, and
when doped it becomes a narrow band gap semiconductor [2]. Sr4Nb4O14 shows
temperature independent thermal conductivity with a value as low as ~ 1 W/m.K for
textured polycrystalline ceramic in the b direction [3]. La doping in Sr4Nb4O14
decreases thermal conductivity due to the higher molecular weight of La. The
maximum solubility of La in single crystal Sr4Nb4O14 is reported to be x= 0.8 [4].
There has been no reported investigation of the effect of non-stoichiometry and
oxidation-reduction on its phase stability and microstructure. In the current research,
Sr4-xLaxNb4O14+
±δ
compositions were synthesised by solid state reaction and spark
plasma sintering and their microstructure and phase contrast were studied.
Page 96
Chapter V. Characterization of Sr4Nb4O14±δ
5.2
EXPERIMENTAL DETAILS
Sr4-xLaxNb4O14+
±δ
(where x=0, 0.2, 0.4, 0.6 and 0.8) were prepared
by solid state reaction. The starting materials were La2O3 (99.99% purity, Sigma
Aldrich), SrCO3 (99.9% purity, Sigma Aldrich) and Nb2O5 (99.9% purity, Alfa
aesar). These powders were mixed in stoichiometric ratios and ball milled using
ethanol as the milling medium in a planetary ball mill for 2 hours. Mixed powders
were calcined at 1250oC for 4 hours. After calcination, the powders were remilled
for 4 hours to break the agglomerates and reduce the particle size.
The powders were sintered using SPS at 1350oC under 50 MPa of pressure
for 5 minutes. A heating rate of 100oC/min was employed in all cases. The sintered
discs were subsequently air annealed and reduced as discussed in the previous
chapter. The bulk density as measured by the Archimedes principle of all sintered
samples was more than 97%.
Page 97
Chapter V. Characterization of Sr4Nb4O14±δ
5.3
RESULTS AND DISCUSSIONS
Table 5.1 presents the theoretical and relative densities of pure and
Sr4-xLaxNb4O14±δ after SPS. The density of all the ceramics increased after
substitution due to the higher atomic weight of La (~ 138.9 amu). All the ceramics
were more than 97% dense.
Table 5.1: Theoretical and relative densities of Sr4-xLaxNb4O14+
±δ ceramics
Theoretical Density Relative Density (%)
(g cm-3)
Sr4Nb4O14±δ
5.21
97.7
Sr3.8La0.2Nb4O14.1±δ 5.27
97.1
Sr3.6La0.4Nb4O14.2±δ 5.34
97.7
Sr3.4La0.6Nb4O14.3±δ 5.40
97.2
Sr3.2La0.8Nb4O14.4±δ 5.47
96.9
Figure 5.1 shows the X ray diffraction patterns for all Sr4-xLaxNb4O14+
±δ
compositions in powder, as sintered, air annealed and reduced conditions. The peaks
matched with Sr4Nb4O14, n=4 (PDF card # 28-1246). After x=0.2, a secondary phase
was produced and indexed as LaNbO4 (PDF card # 01-081-1973) which is a n = 2
PLS phase. The second phase was produced because of the reaction between La and
Nb during calcination as reported by Istomin et al. for the similar compound
(Sr1-xLaxNb2O6) [5]. This gives an indication that the solubility limit of La in
Sr4-xLaxNb4O14+
±δ
is less than x=0.4 in the case of polycrystalline ceramics
synthesised by conventional solid state reaction.
Page 98
Chapter V. Characterization of Sr4Nb4O14±δ
a)
Sr3.2La0.8Nb4O14.2±δ
Sr3.4La0.6Nb4O14.2±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.8La0.2Nb4O14.2±δ
Sr4Nb4O14±δ
Sr4Nb4O14
LaNbO4
b)
Sr3.2La0.8Nb4O14.2±δ
Sr3.4La0.6Nb4O14.2±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.8La0.2Nb4O14.2±δ
Sr4Nb4O14±δ
Sr4Nb4O14
LaNbO4
Page 99
Chapter V. Characterization of Sr4Nb4O14±δ
c)
Sr3.2La0.8Nb4O14.2±δ
Sr3.4La0.6Nb4O14.2±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.8La0.2Nb4O14.2±δ
Sr4Nb4O14±δ
Sr4Nb4O14
LaNbO4
d)
Sr3.2La0.8Nb4O14.2±δ
Sr3.4La0.6Nb4O14.2±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.8La0.2Nb4O14.2±δ
Sr4Nb4O14±δ
Sr4Nb4O14
LaNbO4
Figure 5.1: XRD patterns of Sr4-xLaxNb4O14+
±δ:
(a) powder; (b) sintered; (c) air annealed; (d)
reduced samples
Page 100
Chapter V. Characterization of Sr4Nb4O14±δ
The amount of LaNbO4 was calculated by the RIR method and is given in Table 5.2
Table 5.2: Amount of 2 layer PLS compound in Sr4-xLaxNb4O14+
±δ
ceramics
Powder
Sintered
Air
Reduced
(mass %)
(mass %)
Annealed
%)
(mass
(mass %)
Sr3.8La0.2Nb4O14.1±δ -
-
-
-
Sr3.6La0.4Nb4O14.2±δ -
-
-
-
Sr3.4La0.6Nb4O14.3±δ 3.5
1.5
1.5
0.9
Sr3.2La0.8Nb4O14.4±δ 6.4
2.2
2.4
1.5
Figure 5.2 shows SEM images of polished and then thermally etched
Sr4-xLaxNb4O14+
±δ
ceramic surfaces after sintering by SPS at 1350oC under 50 MPa
pressure. Pure Sr4Nb4O14 has elongated and plate-like grains as reported in the
literature [1]. Upon substitution of La for Sr, the grains become less plate-like as
shown in Figure 5.2(c).
Page 101
Chapter V. Characterization of Sr4Nb4O14±δ
a)
b)
c)
Figure 5.2: SEM images of Sr4-xLaxNb4O14+
±δ
ceramics after polishing and etching: (a)
Sr4Nb4O14±δ; (b) Sr3.8La0.2Nb4O14.1± δ and; (c) Sr3.2La0.8Nb4O14.4± δ
Figure 5.3(a) shows a typical bright field TEM image of Sr4Nb4O14 based
ceramic after air annealing. The typical grain size of the ceramic was ~ 0.3 µm and
most of the grains appear defect free. Figure 5.3(b) shows a high resolution image of
(020) lattice planes. The inter-planar spacing corresponds to the 4 layer Sr4Nb4O14
(~ 12.6 Å). Figure 5.3(c) shows the corresponding diffraction patter. Figure 5.3 (d)
shows a bright field image of Sr3.2La0.8Nb4O14.2±δ ceramic which gives an overall
view of the ceramics microstructure. From the XRD data, generation of the 2 layered
PLS compound LaNbO4 was confirmed and is marked in Figure 5.3(e) as separate
single phase grains. Figure 5.3(f) shows a magnified image of a separate single grain
Page 102
Chapter V. Characterization of Sr4Nb4O14±δ
of LaNbO4. In some areas, LaNbO4 was uniformly distributed while in other, it was
found in clusters in the microstructure as shown in Figure 5.3 (d and e).
a)
b)
c)
d)
e)
f)
Figure 5.3: TEM images of: (a) air annealed Sr4Nb4O14±δ; (b) high magnification image of air
annealed Sr4Nb4O14±δ; (c) corresponding diffraction pattern for Sr4Nb4O14±δ; (d) air annealed
Sr3.2La0.8Nb4O14.2±δ; (e) air annealed Sr3.2La0.8Nb4O14.2±δ showing regions of Sr4Nb4O14 and LaNbO4;
(f) high magnification image of a single LaNbO4 grain
EDS analysis was done on three regions marked as 1, 2 and 3 in Figure
5.3(e). EDS results also confirmed the Sr/Nb (region 1) and La/Nb (region 3) rich
Page 103
Chapter V. Characterization of Sr4Nb4O14±δ
phases and the estimated atomic percentage is given in the Table 5.3. These rough
estimates gives an indication of the two separate phases presents inside the structure
namely Sr4Nb4O14 and LaNbO4.
Table 5.3: Percentage mass amount of different elements (excluding oxygen) in
Sr4-xLaxNb4O14+
±δ
ceramics
Sr (%)
Nb (%)
La (%)
Total
Region 1
31.19
34.67
5.09
100.00
Region 2
30.89
37.09
10.96
100.00
Region 3
6.01
27.08
42.77
100.00
Figure 5.4(a-b) shows bright filed TEM image of reduced ceramics. After
reduction, the density of Sr4Nb4O14 decreased slightly (from 5.09 g cm-3to 5.03 g cm3
) and a few planar defects were generated. These results are consistent with XRD
patterns which indicate a slight increase in FWHM of the reduced ceramics. Figure
5.4(b) shows a high resolution image of a typical planar defect in reduced Sr4Nb4O14.
Figure 5.4(c) shows an image of Sr3.6La0.4Nb4O14.2±δ after reduction. The density of
Sr3.6La0.4Nb4O14+
±δ
decreased slightly (from 5.3 g cm-3 to 5.18 g cm-3) after
reduction and the amount of secondary phase stayed nearly the same. The clusters of
Sr4Nb4O14 and LaNbO4 were also observed in the reduced ceramic as shown in
Figure 5.4(d and e). Figure 5.4(e) shows a magnified Sr/Nb rich region with planar
defects inside the microstructure. It is anticipated that the LaNbO4 has started to
form and was distributed throughout the grain. Most of the LaNbO4 grains were
plain without any planar defects, but some had planar defects as La was substituted
Page 104
Chapter V. Characterization of Sr4Nb4O14±δ
in Sr4Nb4O14 and shown in Figure 5.4(d). The shape and size of the grain is similar
to those observed in section 4.5. It was not possible to determine the phase which
was formed at the region of the planar defect due to the complex nature of the
defects as discussed in the case of Nb and Ta doped La4Ti4O14 (section 4.5).
b)
a)
c)
d)
e)
Figure 5.4: TEM images of: (a) reduced Sr4Nb4O14±δ; (b) high magnification image of reduced
Sr4Nb4O14±δ; (c-d) typical regions of reduced Sr3.2La0.8Nb4O14.4±δ; (e) high magnification image of a
typical Sr/Nb rich region
Page 105
Chapter V. Characterization of Sr4Nb4O14±δ
Figure 5.5(a–c) shows the XPS spectra of Sr 3d, Nb 3d, O 1s and La 3d
surface electrons in Sr4Nb4O14 based ceramics without ion beam etching. Figure
5.4(a) shows the XPS spectra of typical Sr 3d doublet. For air annealed Sr4Nb4O14
and Sr3.6La0.4Nb4O14.2±δ, the intensity and peak position was the same and given in
Table 5.4. This gives an indication of the stability of Sr in pure and doped ceramics,
the binding energy values for Sr 3d5/2 and Sr 3d3/2 were 132.8 eV and 134.4 eV
respectively. This matches well with the values reported in literature [6]. Figure
5.4(b) shows the XPS spectra of Nb 3d electrons of pure and La substituted
Sr4Nb4O14. The binding energy values for Nb 3d showed a clear doublet Nb 3d5/2 and
Nb 3d3/2 which is a characteristic of Nb containing oxide compounds. The binding
energy values for Nb 3d5/2 and Nb 3d3/2 in pure Sr4Nb4O14 were ~ 206.6 eV and ~
209.5 eV respectively which were similar to those in La substituted Sr4Nb4O14
(Table 5.4). The difference between the two peaks was 2.8 eV. This value is in good
agreement with the one published in literature [6]. The O 1s core level spectra are
shown in Figure 5.5(c). Both spectra had two convoluted peaks. The component at
530.0 eV gives the main O 1s state of the oxygen in Sr4Nb4O14 [7]. The higher
binding energy component at 531.4 eV might indicate chemisorbed oxygen or
absorbed hydroxyl ions. After La substitution, the peak position was at 529.8 eV.
Page 106
Chapter V. Characterization of Sr4Nb4O14±δ
a)
Sr3.6La0.4Nb4O14.2±δ
b)
Sr3.6La0.4Nb4O14.2±δ
Page 107
Chapter V. Characterization of Sr4Nb4O14±δ
c)
Sr3.6La0.4Nb4O14.2±δ
d)
Sr3.6La0.4Nb4O14.2±δ
air annealed
Figure 5.5: XPS spectra of Sr3.6La0.4Nb4O14.2±δ: (a) Sr 3d; (b) Nb 3d; (c) O1s and; (d) La3d
The ∆BE values were also calculated and are presented in the Table 5.4. The
∆BE values for (Sr-O) in the case of pure and La substituted Sr4Nb4O14 was 397.2
eV and 397 eV respectively. The decrease in binding energy was caused by the
Page 108
Chapter V. Characterization of Sr4Nb4O14±δ
weakening of the bonds as a result of the second phase. These values matched that
for Sr substituted La4Ti4O14 (397 eV). For ∆BE (Nb-O) the values were 323.2 eV
and 323.1 eV for pure and La substituted Sr4Nb4O14 respectively. These values are in
good agreement with the literature [6]. For ∆BE (La-O) the value was 305.2 eV. This
value was higher than that measured for La4Ti4O14 (304.7 eV). Since La has gone to
make LaNbO4 and the ratio of La:O in LaNbO4 is 1:4 which is higher than the one
for La4Ti4O14 which is 1:3.5, this explains the increase in the ∆BE (La-O) in La
substituted Sr4Nb4O14.
Table 5.4: XPS data of Sr3.2La0.4Nb4O14.2±δ ceramics
Material
Peak Position (eV)
Sr3d
Sr4Nb4O14 air
Nb3d
O1s
Binding Energy Difference
(∆eV)
Sr-O
Nb-O
La-O
Sr3d3/2
Sr3d5/2 Nb3d3/2 Nb3d5/2
134.4
132.8
209.5
206.8
530.0
397.2
323.2
-
134.4
132.8
209.4
206.7
529.8
397.0
323.1
305.2
annealed
Sr3.2La0.4Nb4O14.2±δ
air annealed
Page 109
Chapter V. Characterization of Sr4Nb4O14±δ
5.4
CONCLUSION
Pure and La substituted Sr4Nb4O14 were prepared by solid state reaction and
spark plasma sintering. The limit of La substitution in polycrystalline Sr4Nb4O14 is
found to be x< 0.4. La substitution produced a secondary phase LaNbO4 when the
level of substitution increased to x=0.4. The amount of secondary phase decreased
after sintering and reduction. La substitution also changed the size and shape of the
grains and they became less plate-like. The main mechanism of La accommodation
inside the structure is by the formation of the 2 layer PLS compound, LaNbO4.
Page 110
Chapter V. Characterization of Sr4Nb4O14±δ
REFERENCES
[1] H. Ning, H. Yan, M.J. Reece, Journal of the American Ceramic Society, 93 [5]
(2010) 1409.
[2] A. Sakai, T. Kanno, K. Takahashi, Y. Yamada, H. Adachi, Journal of Applied
Physics, 108 (2010) 103706.
[3] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
[4] A. Sakai, T. Kanno, K. Takahashi, A. Omote, H. Adachi, Y. Yamada, X.D. Zhou,
Journal of the American Ceramic Society, 95 (2012) 1750-1755.
[5] S.Y. Istomin, O.G. Dyachenko, E.V. Antipov, G. Svensson, M. Nygren,
Materials Research Bulletin, 29 (1994) 743-749.
[6] V.V. Atuchin, J.C. Grivel, A.S. Korotkov, Z. Zhang, Journal of Solid State
Chemistry, 181 (2008) 1285-1291.
[7] Z. Gao, H. Ning, C. Chen, R. Wilson, B. Shi, H. Ye, H. Yan, M.J. Reece, J.L.
Jones, Journal of the American Ceramic Society, 96 (2013) 1163-1170.
Page 111
Chapter VI. Thermal Conductivity of PLS Compounds
Chapter VI. Thermal Conductivity of PLS
Compounds
6.1
INTRODUCTION
The aim of this chapter is to investigate the effect of acceptor-donor
substitution and compositional non stoichiometry on the thermal conductivity of PLS
compounds. Currently there is a limited amount of literature relating to the effect of
substituting elements and compositional non stoichiometry on the thermal
conductivity of PLS ceramics [1-3]. The materials investigated for thermal
conductivity were La4Ti4O14 and Sr4Nb4O14. There are no reported data for the
thermal conductivity of La4Ti4O14. Liu et al. calculated minimum thermal
conductivity of pyrochlore La2Ti2O7 to be 1.27 W/m.K using Clarke’s relation which
will be discussed later [4]. For polycrystalline Sr4Nb4O14, thermal conductivity is
temperature independent (above room temperature) and has a value of ~ 1-2 W/m.K
[2]. In this work, the effect of substitution and reduction on thermal conductivity is
discussed in relation to the microstructures for La4Ti4O14 and Sr4Nb4O14 which were
discussed in chapters 4 and 5.
Page 112
Chapter VI. Thermal Conductivity of PLS Compounds
6.2
EXPERIMENTAL DETAILS
The samples for thermal characterization were prepared by a mixed oxide
route and SPS as described in sections 4.2 and 5.2. A Netszch LFA 457 microflash
was employed to measure the thermal diffusivity which was then used to calculate
thermal conductivity using equation 3.3. All the measurements were carried out
under Ar environment. The sintered discs were ground flat but left unpolished to
minimise laser reflectance. A Kontact-chemi Graphite 33 spray was used to coat
both surfaces of the sample to maximise the absorption of the laser light.
Measurements were taken at room temperature and then at intervals of 100 oC up to
a temperature of 800 oC. Three measurements were taken at each temperature to
minimize the scatter in the data. In order to ensure the thermal stability across the
sample at each temperature, a period of 2 minutes between the laser shots was used.
Temperature fluctuation was kept to a minimum of ± 1 oC for all measurements. The
thermal diffusivity measurement was repeated to check reproducibility of the
measurements and materials.
The specific heat capacity (Cp) was measured by Netzsch STA (449 F3
Jupiter®) under N2 environment. A square sample with dimensions 4mm x 4mm x
1mm was cut from the sintered disc and ground flat. The instrument was checked for
calibration by running a standard sapphire test sample and comparing the values with
the standard values. The Cp data below 300 oC (573 K) were not reliable (due to
non-availability of calibration for the equipment below 300 oC) and they were not
used in the calculations. All the results obtained were reproducible and there was no
obvious colour change of the oxidized or reduced samples after the measurements.
Page 113
Chapter VI. Thermal Conductivity of PLS Compounds
RESULTS
6.3
La4Ti4O14
Chapter 4 reports a detailed microstructural examination of A (acceptor) and
B (donor) site substituted La4Ti4O14. After substitution, the crystal structure of
La4Ti4O14 accommodates the deviation from the stoichiometric composition by
incorporating planar defects in the structure. Since these planar defects are of the
order of few angstroms (typically 3 Å) it is expected to see a reduction in thermal
conductivity after hydrogen reduction and/or substitution in La4Ti4O14 ceramic.
6.3.1
La4-xSrxTi4O14-
±δ
Figure 6.1 shows the thermal diffusivity of air annealed and reduced
La4-xSrxTi4O14-
±δ
ceramics. The thermal diffusivity of all the ceramics showed near
temperature independency. The thermal diffusivity increased with increasing Sr
substitution after air annealing and reduction. Compared to air annealed ceramics,
the corresponding thermal diffusivity of all the ceramics decreased after reduction.
This decrease was due to the increased density of the nanoscale intergrowth of 5
layer PLS structure inside the 4 layer PLS structure (Figure 4.4).
Page 114
Chapter VI. Thermal Conductivity of PLS Compounds
La4Ti4O14±δ
La3.9Sr0.1Ti4O13.9±δ
La3.8Sr0.2Ti4O13.8±δ
a)
La3.7Sr0.3Ti4O13.7±δ
La3.6Sr0.4Ti4O13.6±δ
b)
La4Ti4O14±δ
La3.9Sr0.1Ti4O13.9±δ
La3.8Sr0.2Ti4O13.8±δ
La3.7Sr0.3Ti4O13.7±δ
La3.6Sr0.4Ti4O13.6±δ
Figure 6.1: Variation in thermal diffusivity of La4-xSrxTi4O14-
±δ
ceramics: (a) air annealed;
(b) reduced
Figure 6.2 compares the Cp measured and reported value of sapphire
standard. The measured values are less than the reported values and the error in the
measurement was calculated to be ~ 7%.
Page 115
Chapter VI. Thermal Conductivity of PLS Compounds
Sapphire Value from Literature
Measured Sapphire value
Figure 6.2: Variation in specific heat capacity of sapphire with temperature
Figure 6.3 shows the specific heat capacity of the La4-xSrxTi4O14-
±δ
ceramics.
The heat capacity increases with increasing temperature and reaches a maximum
value. Heat capacity of air annealed La4Ti4O14±δ is higher than the reduced
La4Ti4O14±δ. This decrease in specific heat capacity in reduced samples can be
explained by the fact that due to the removal of oxygen by reduction, the binding
energy decreased as observed in the section 4.3. After Sr doping, the binding energy
is still lower than for the pure air annealed La4Ti4O14±δ which also contributes for the
decrease in the specific heat of Sr substituted La4Ti4O14 [5].
Page 116
Chapter VI. Thermal Conductivity of PLS Compounds
a)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
b)
La3.2Sr0.8Ti4O13.6±δ Air annealed
La3.2Sr0.8Ti4O13.6±δ Reduced
Figure 6.3: Variation in specific heat capacity of: (a) La4Ti4O14±δ; (b) La3.2Sr0.8Ti4O13.6±δ
ceramics with temperature
Page 117
Chapter VI. Thermal Conductivity of PLS Compounds
Figure 6.4 compares the specific heat capacity of reduced La4Ti4O14±δ
ceramics. The ceramic was rerun to check the reproducibility of the results and the
results were within the experimental error.
Figure 6.4: Variation in specific heat capacity of reduced La4Ti4O14±δ with temperature
Table 6.1 gives the densities of La4-xSrxTi4O14-
±δ
ceramics after air annealing
and after hydrogen reduction. All the ceramics were more than 98% dense and due to
the lower atomic weight of Sr (87.6 amu) compared to La (138.9 amu), the
theoretical density of all the Sr substituted ceramics decreased. The density of the
ceramics after reduction is lower than the air annealed ceramics due to the removal
of oxygen.
Page 118
Chapter VI. Thermal Conductivity of PLS Compounds
Table 6.1: Change in measured densities after Sr substitution and hydrogen reduction
Sample
Density
after
air Density after reduction
annealing (g/ cm-3)
(g/ cm-3)
La4Ti4O14±δ
5.74±0.04
5.62±0.01
La3.8Sr0.2Ti4O13.9±δ
5.68±0.08
5.59±0.03
La3.6Sr0.4Ti4O13.8±δ
5.62±0.05
5.57±0.04
La3.4Sr0.6Ti4O13.7±δ
5.59±0.05
5.52±0.03
La3.2Sr0.8Ti4O13.6±δ
5.53±0.03
5.47±0.01
Figure 6.5 (a) shows the thermal conductivity of air annealed ceramics
estimated using equation 3.3. The thermal conductivity of La4Ti4O14±δ is nearly
temperature independent as reported in the literature for similar PLS compounds
e.g.Sr4Nb4O14 [2]. The value of thermal conductivity for La4Ti4O14±δ is ~ 1.2 W/m.K
in this work, which is close to that of other layered structured compounds, e.g.
Bi4Ti3O12 [6]. After Sr substitution, the value of thermal conductivity increases with
increasing temperature. Figure 6.5(b) shows the thermal conductivity of reduced
La4-xSrxTi4O14-
±δ
ceramics. After reduction, the values of thermal conductivity
decrease as compared to the air annealed samples and the value is less temperature
dependent. Figure 6.5(c) and (d) compares the thermal conductivity values of air
annealed and reduced ceramics.
Page 119
Chapter VI. Thermal Conductivity of PLS Compounds
La4Ti4O14±δ
La3.9Sr0.1Ti4O13.9±δ
a)
La3.8Sr0.2Ti4O13.8±δ
La3.7Sr0.3Ti4O13.7±δ
La3.6Sr0.4Ti4O13.6±δ
La4Ti4O14±δ
b)
La3.9Sr0.1Ti4O13.9±δ
La3.8Sr0.2Ti4O13.8±δ
La3.7Sr0.3Ti4O13.7±δ
La3.6Sr0.4Ti4O13.6±δ
Page 120
Chapter VI. Thermal Conductivity of PLS Compounds
c)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
d)
La3.2Sr0.8Ti4O13.6±δ Air annealed
La3.2Sr0.8Ti4O13.6±δ Reduced
Figure 6.5: Variation in thermal conductivity of: (a) air annealed La4-xSrxTi4O14- ±δ; (b) reduced
La4-xSrxTi4O14-
±δ;
(c) La4Ti4O14±δ; (d) La3.2Sr0.8Ti4O13.6±δ with temperature
Figure 6.6 shows the variation of thermal conductivity with the amount of Sr
at 573 K (300 oC). The thermal conductivity decreases with increasing Sr content
due to the generation of nanoscale intergrowths.
Page 121
Chapter VI. Thermal Conductivity of PLS Compounds
Figure 6.6: Variation of thermal conductivityof reduced La4-xSrxTi4O14-
±δ
with Sr content at
573 K
Page 122
Chapter VI. Thermal Conductivity of PLS Compounds
6.3.2
La4Ti4-xTaxO14+
±δ
The thermal diffusivity of air annealed and reduced La4Ti4-xTaxO14+
±δ
ceramics are given in Figure 6.7. The thermal diffusivity of all the ceramics is nearly
temperature independent and decreased with increasing Ta addition. After reduction,
the thermal diffusivity of the ceramics decreased with increasing temperature.
a)
La4Ti4O14±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.4Ta0.6O14.3±δ
b)
La4Ti4O14±δ
La4Ti3.9Ta0.1O14.05±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.4Ta0.6O14.3±δ
Figure 6.7: Variation in thermal diffusivity of La4Ti4-xTaxO14+
±δ
ceramics: (a) air annealed;
(b) reduced
Page 123
Chapter VI. Thermal Conductivity of PLS Compounds
The specific heat capacity of La4Ti4-xTaxO14+
±δ
ceramics (x=0 and 0.6) are
given in Figure 6.8. The heat capacity increases with increasing temperature and
reaches a maximum value. The heat capacity for La4Ti4-xTaxO14+
±δ
ceramics is
lower than the pure La4Ti4O14. According to the Dulong Petit approximation, the
product of specific heat capacity and molecular weight is roughly equal to 3R, where
R is a general gas constant [7]. This means that increasing the atomic weight would
decrease the specific heat capacity of a solid. The atomic weight of
La4Ti4-xTaxO14+
±δ
ceramics increased as a result of Ta substitution in La4Ti4O14.
This increase in the atomic weight explains decrease in the specific heat capacity of
substituted ceramics.
a)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
b)
La4Ti3.4Ta0.6O14.3±δ Air annealed
La4Ti3.4Ta0.6O14.3±δ Reduced
Figure 6.8: Variation in specific heat capacity of: (a) La4Ti4O14±δ; (b) La4Ti3.4Ta0.6O14.3±δ ceramics
with temperature
Page 124
Chapter VI. Thermal Conductivity of PLS Compounds
Table 6.2 gives the densities of La4Ti4-xTaxO14+
±δ
ceramics. All the ceramics
were more than 97% dense and due to the higher molecular weight of Ta than Ti, the
theoretical density of all the Ta substituted ceramics increased. The density of the
reduced ceramics is lower than the air annealed ceramics due to the removal of
oxygen.
Table 6.2: Change in measured densities after Ta substitution and hydrogen reduction
Sample
Density
after
air Density after reduction
annealing (g cm-3)
(g cm-3)
La4Ti4O14±δ
5.74±0.04
5.62±0.01
La4Ti3.9Ta0.1O14.05±δ
5.84±0.01
5.77±0.04
La4Ti3.8Ta0.2O14.1±δ
5.91±0.02
5.84±0.02
La4Ti3.6Ta0.4O14.2±δ
5.95±0.04
5.90±0.03
La4Ti3.4Ta0.6O14.3±δ
6.07±0.02
5.98±0.03
The thermal conductivity of air annealed ceramics is given in Figure 6.9(a).
The thermal conductivity of La4Ti4-xTaxO14+
±δ
ceramics is nearly temperature
independent. The thermal conductivity decreased with increasing Ta content. Figure
6.9(b) shows the thermal conductivity of reduced La4Ti4-xTaxO14+
±δ
ceramics. After
reduction, the value of thermal conductivity decreased as compared to the air
annealed samples. The decrease in thermal conductivity after reduction was less
compared to La4-xSrxTi4O14-
±δ
ceramics. Figure 6.9(c) and (d) compares the thermal
conductivity values of air annealed and reduced ceramics.
Page 125
Chapter VI. Thermal Conductivity of PLS Compounds
La4Ti4O14±δ
La4Ti3.9Ta0.1O14.05±δ
a)
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.4Ta0.6O14.3±δ
La4Ti4O14±δ
b)
La4Ti3.9Ta0.1O14.05±δ
La4Ti3.8Ta0.2O14.1±δ
La4Ti3.6Ta0.4O14.2±δ
La4Ti3.4Ta0.6O14.3±δ
Page 126
Chapter VI. Thermal Conductivity of PLS Compounds
c)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
d)
La4Ti3.4Ta0.6O14.3±δ Air annealed
La4Ti3.4Ta0.6O14.3±δ Reduced
Figure 6.9: Variation in thermal conductivity of: (a) air annealed La4Ti4-xTaxO14+ ±δ; (b) reduced
La4Ti4-xTaxO14+
; (c) La4Ti4O14±δ; (d) La4Ti3.4Ta0.6O14.3±δ with temperature
±δ
Figure 6.10 shows the data for variation of thermal conductivity with the
amount of Ta at 573 K. There is a decreasing trend in thermal conductivity with
increasing Ta content due to the higher molecular weight of Ta than Ti and
nanoscale intergrowths.
Page 127
Chapter VI. Thermal Conductivity of PLS Compounds
Figure 6.10: Variation of thermal conductivity with Ta content for reduced La4Ti4-xTaxO14+
±δ
at
573 K
Page 128
Chapter VI. Thermal Conductivity of PLS Compounds
6.3.3 La4Ti4-xNbxO14+
±δ
The thermal diffusivity data of air annealed and reduced La4Ti4-xNbxO14±δ
ceramics are shown in Figure 6.11. The thermal diffusivity of all the ceramics
showed the same near temperature independency as La4-xSrxTi4O14La4Ti4-xTaxO14+
±δ
±δ
and
ceramics. The thermal diffusivity decreased with increasing Nb
addition both in the case of air annealed and reduced ceramics. The thermal
diffusivity value of reduced La4Ti4-xNbxO14+
±δ
ceramics is lower than the air
annealed ceramics.
La4Ti4O14±δ
a)
La4Ti3.8Nb0.2O14.1±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.4Nb0.6O14.3±δ
La4Ti4O14±δ
b)
La4Ti3.8Nb0.2O14.1±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.4Nb0.6O14.3±δ
Figure 6.11: Variation in thermal diffusivity of La4Ti4-xNbxO14
±δ
ceramics: (a) air annealed;
(b) reduced
Page 129
Chapter VI. Thermal Conductivity of PLS Compounds
Figure 6.12 shows the specific heat capacity of La4Ti4-xNbxO14+
±δ
ceramics.
The specific heat capacity increases with increasing temperature and reaches a
maximum value. The specific heat capacity decreases with increasing substitution
which is according to the Dulong Petite approximation as discussed previously
(section 6.3.2).
a)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
b)
La4Ti3.4Nb0.6O14.3±δ Air annealed
La4Ti3.4Nb0.6O14.3±δ Reduced
Figure 6.12: Variation in specific heat capacity of: (a) La4Ti4O14±δ; (b) La4Ti3.4Nb0.6O14.3±δ
ceramics with temperature
Table 6.3 gives the densities of air annealed and reduced La4Ti4-xNbxO14+
±δ
ceramics. All the ceramics were more than 98% dense and due to the higher
Page 130
Chapter VI. Thermal Conductivity of PLS Compounds
molecular weight of Nb (92.9 amu) compared to Ti (47.87 amu), the density of the
Nb substituted ceramics increased. The density of the ceramics after hydrogen
reduction is lower than the air annealed ceramics due to the loss of oxygen.
Table 6.3: Change in measured densities after Nb substitution and hydrogen reduction
Sample
Density
after
air Density after reduction
annealing (g cm-3)
(g cm-3)
La4Ti4O14±δ
5.74±0.04
5.62±0.01
La4Ti3.8Nb0.2O14.1±δ
5.82±0.04
5.68±0.04
La4Ti3.6Nb0.4O14.2±δ
5.87±0.01
5.75±0.04
La4Ti3.4Nb0.6O14.3±δ
5.88±0.04
5.76±0.06
Figure 6.13(a) shows thermal conductivity of air annealed ceramics. The
thermal conductivity value of La4Ti4-xNbxO14+
±δ
ceramics is nearly temperature
independent. The value of thermal conductivity decreased with increasing Nb
content. Figure 6.13(b) shows thermal conductivity of reduced La4Ti4-xNbxO14+
±δ
ceramics. After reduction, the value of thermal conductivity decreases as compared
to the air annealed samples. Figure 6.13(c) and (d) compares thermal conductivity
values of air annealed and reduced ceramics.
Page 131
Chapter VI. Thermal Conductivity of PLS Compounds
a)
La4Ti4O14±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.4Nb0.6O14.3±δ
b)
La4Ti4O14±δ
La4Ti3.8Nb0.2O14.1±δ
La4Ti3.6Nb0.4O14.2±δ
La4Ti3.4Nb0.6O14.3±δ
Page 132
Chapter VI. Thermal Conductivity of PLS Compounds
c)
La4Ti4O14±δ Air annealed
La4Ti4O14±δ Reduced
d)
La4Ti3.4Nb0.6O14.3±δ Air annealed
La4Ti3.4Nb0.6O14.3±δ Reduced
Figure 6.13: Variation in thermal conductivity of: (a) air annealed La4Ti4-xNbxO14+ ±δ; (b) reduced
La4Ti4-xNbxO14+
; (c) La4Ti4O14±δ; (d) La4Ti3.4Nb0.6O14.3±δ with temperature
±δ
The variation of thermal conductivity with the increasing amount of Nb at
573 K is given in Figure 6.14. The thermal conductivity decreases with increasing
Nb content due to the higher atomic weight of Nb (~ 92.9 amu) than Ti (~ 47.8 amu).
Also, the amount of secondary phase (LaNbO4) increases with increasing Nb
Page 133
Chapter VI. Thermal Conductivity of PLS Compounds
content, it helps in scattering phonons and thus decreasing thermal conductivity
value.
Figure 6.14: Variation of thermal conductivity with Nb content for reduced
La4Ti4-xNbxO14+
±δ
at 573 K
Page 134
Chapter VI. Thermal Conductivity of PLS Compounds
6.4
Sr4-xLaxNb4O14+ ±δ
Thermal properties of A site substituted Sr4Nb4O14 will be discussed in this
section. After substitution of Sr with La, the crystal structure of Sr4Nb4O14
accommodates the non-stoichiometry by forming a second phase which is 2 layer
PLS compound LaNbO4 (section 5.3).
Figure 6.15 shows the thermal diffusivity of air annealed and reduced
Sr4-xLaxNb4O14+
±δ
ceramics. The thermal diffusivity of all the ceramics showed the
same near temperature independency as previously reported [2]. The thermal
diffusivity decreased with increasing La addition. This thermal diffusivity value is
lower than the one reported by Sparks et al. for polycrystalline Sr4Nb4O14 [2].
Page 135
Chapter VI. Thermal Conductivity of PLS Compounds
a)
Sr4Nb4O14±δ
Sr3.8La0.2Nb4O14.1±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.4La0.6Nb4O14.3±δ
Sr3.2La0.8Nb4O14.4±δ
b)
Sr4Nb4O14±δ
Sr3.8La0.2Nb4O14.1±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.4La0.6Nb4O14.3±δ
Sr3.2La0.8Nb4O14.4±δ
Figure 6.15: Variation in thermal diffusivity of Sr4-xLaxNb4O14+
±δ
ceramics: (a) air
annealed; (b) reduced
The heat capacity of Sr4-xLaxNb4O14+
±δ
ceramics increases with increasing
temperature as given in Figure 6.16 and reaches a maximum value. The heat capacity
decreases with increasing La content, but after reduction, the heat capacity of all the
ceramics were similar. Referring back to the Dulong Petit law, the specific heat
Page 136
Chapter VI. Thermal Conductivity of PLS Compounds
varies inversely to formula weight. Since La substitution for Sr increased the total
formula weight of the ceramics, this in turn lowered the specific heat capacity value.
a)
Sr4Nb4O14±δ Air annealed
Sr4Nb4O14±δ Reduced
b)
Sr3.2La0.8Nb4O14.4±δ Air annealed
Sr3.2La0.8Nb4O14.4±δ Reduced
Figure 6.16: Variation in specific heat capacity of: (a) Sr4Nb4O14±δ; (b) Sr3.2La0.8Nb4O14.4±δ
ceramics with temperature
Table 6.4 gives the densities of Sr4-xLaxNb4O14+
±δ
ceramics after air
annealing and after reduction. All the ceramics were more than 98% dense and due
to the higher molecular weight of La (138.9 amu) than Sr (87.6amu), the density of
Page 137
Chapter VI. Thermal Conductivity of PLS Compounds
La substituted ceramics increased. The density of the ceramics decreased after
reduction.
Table 6.4: Change in measured densities after La substitution and hydrogen reduction
Sample
Density
after
air Density after reduction
annealing (g cm-3)
(g cm-3)
Sr4Nb4O14±δ
5.09± 0.01
5.03±0.02
Sr3.8La0.2Nb4O14.1±δ
5.12±0.03
5.07±0.02
Sr3.6La0.4Nb4O14.2±δ
5.22±0.02
5.19±0.01
Sr3.4La0.6Nb4O14.3±δ
5.25±0.01
5.16±0.04
Sr3.2La0.8Nb4O14.4±δ
5.30±0.01
5.18±0.03
Figure 6.17(a) shows the thermal conductivity of air annealed ceramics. The
thermal conductivity of Sr4-xLaxNb4O14+
substitution.
Figure
Sr4-xLaxNb4O14+
±δ
6.17(b)
shows
±δ
ceramics decreased with increasing La
thermal
conductivity
of
reduced
ceramics. After reduction, the value of thermal conductivity
decreased until x= 0.2 and then increased. Figure 6.17(c) and (d) compares thermal
conductivity values of air annealed and reduced Sr4-xLaxNb4O14+
±δ
(x=0 and 0.8)
ceramics. For Sr4Nb4O14, the thermal conductivity decreases after reduction. This
decrease is consistent with the pure La4Ti4O14 in which thermal conductivity
decreases after reduction due to the generation of planar defects. For
Sr4-xLaxNb4O14+
±δ,
the thermal conductivity increases after reduction. This increase
is related to the decreasing amount of secondary phase LaNbO4 (Table 5.2) which
resulted in less phonon scattering and thus increasing thermal conductivity.
Page 138
Chapter VI. Thermal Conductivity of PLS Compounds
Sr4Nb4O14±δ
Sr3.8La0.2Nb4O14.1±δ
a)
Sr3.6La0.4Nb4O14.2±δ
Sr3.4La0.6Nb4O14.3±δ
Sr3.2La0.8Nb4O14.4±δ
b)
Sr4Nb4O14±δ
Sr3.8La0.2Nb4O14.1±δ
Sr3.6La0.4Nb4O14.2±δ
Sr3.4La0.6Nb4O14.3±δ
Sr3.2La0.8Nb4O14.4±δ
Page 139
Chapter VI. Thermal Conductivity of PLS Compounds
c)
Sr4Nb4O14±δ Air annealed
Sr4Nb4O14±δ Reduced
d)
Sr3.2La0.8Nb4O14.4±δ Air annealed
Sr3.2La0.8Nb4O14.4±δ Reduced
Figure 6.17: Variation in thermal conductivity of: (a) air annealed Sr4-xLaxNb4O14+ ±δ; (b) reduced
Sr4-xLaxNb4O14+
±δ;
(c) Sr4Nb4O14±δ; (d) Sr3.2La0.8Nb4O14.4±δ with temperature
Figure 6.18 shows the variation of thermal conductivity with the amount of
La at 573 K. Thermal conductivity tends to decrease with increasing La content after
air annealing. This decrease is the result of the mass difference caused by the
substitution of La for Sr and the secondary phase (LaNbO4).
Page 140
Chapter VI. Thermal Conductivity of PLS Compounds
Figure 6.18: Variation of thermal conductivity with La content at 573 K: (a) air annealed
and; (b) reduced
6.5
DISCUSSIONS
The thermal conductivities of La4Ti4O14 and Sr4Nb4O14 ceramics follow near
room temperature independency after air annealing and reduction. The thermal
conductivity was in the range of ~ 1.2 W/m.K. After substitution, the thermal
conductivity decreases slightly.
The thermal conductivity of a material depends upon its lattice and electronic
contributions. The electronic contribution of thermal conductivity can be calculated
by the Wiedemann–Franz law and is given by
(2.4)
Where κelec is the electronic contribution of thermal conductivity, L is the
Lorenz number, T is the absolute temperature and σ is the electrical conductivity.
The value of L was taken from literature and it was 2.44x10-8 WΩK-2 at 573 K.
Electrical conductivity was measured using the four probe method under vacuum
Page 141
Chapter VI. Thermal Conductivity of PLS Compounds
with laboratory built equipment [8] and was estimated to be ~ 0.5 S/m for
La3.2Sr0.8Ti4O13.6±δ. This was the highest electrical conductivity measured for the
substituted PLS ceramics in this work. Based on these values, the κelec for La4Ti4O14
and Sr4Nb4O14 based compositions was estimated to be ~ 7x10-9 W/m.K.
In the case of La4Ti4O14, the thermal conductivity of air annealed and reduced
ceramics increased after Sr substitution but decreased after Ta/Nb substitution. This
change in thermal conductivity can be explained by looking at the dependency of
thermal conductivity on average atomic mass. [9]
(6.1)
Where E is the elastic modulus, ρ is the density and Ω is the average atomic
volume and kB is the Boltzmann’s constant. The minimum thermal conductivity
values are presented in Table 6.5.
The average atomic volume for La4Ti4O14 and Sr4Nb4O14 based compositions
was calculated by following relation and tabulated in Table 6.5 [10]
(6.2)
Where M is the average molar mass of the unit cell, m is the number of atoms
per formula unit, ρ is the density and NA is the Avogadro’s number.
From the above two expressions, it is evident that the large mean atomic
mass and low elastic modulus favour low thermal conductivity [9]. The atomic
Page 142
Chapter VI. Thermal Conductivity of PLS Compounds
masses of La4Ti4O14 and Sr4Nb4O14 based compositions are given in table 6.5. (The
atomic weights were calculated based on the formula La2Ti2O7 and Sr2Nb2O7).
Table 6.5: Relative formula weight, mean atomic volume and minimum thermal
conductivity of air annealed La4Ti4O14 and Sr4Nb4O14 based compositions at 973 K
Composition
Formula
Mean
Minimum
Experimental
weight
Atomic
Thermal
Thermal
(amu)
Volume
Conductivity
Conductivity
(nm3)
(W/m.K)
(W/m.K)
La4Ti4O14±δ
485.54
0.01280
1.33
1.31±0.07
La3.8Sr0.2Ti4O13.9±δ
479.61
0.01280
1.35
1.17±0.09
La3.6Sr0.4Ti4O13.8±δ
473.68
0.01280
1.36
1.24±0.07
La3.4Sr0.6Ti4O13.7±δ
467.76
0.01270
1.38
1.25±0.07
La3.2Sr0.8Ti4O13.6±δ
461.83
0.01270
1.38
1.25±0.08
La4Ti3.9Ta0.1O14.05±δ 492.59
0.01270
1.30
1.08±0.08
La4Ti3.8Ta0.2O14.1±δ
499.65
0.01270
1.29
1.03±0.09
La4Ti3.6Ta0.4O14.2±δ
514.55
0.01300
1.24
1.09±0.10
La4Ti3.4Ta0.6O14.3±δ
527.86
0.01310
1.20
1.00±0.09
La4Ti3.8Nb0.2O14.1±δ
490.84
0.01270
1.30
1.13±0.08
La4Ti3.6Nb0.4O14.2±δ
496.15
0.01270
1.28
1.04±0.09
La4Ti3.4Nb0.6O14.3±δ
501.45
0.01280
1.25
0.95±0.09
Sr4Nb4O14±δ
473.04
0.01400
1.19
1.07±0.10
Sr3.8La0.2Nb4O14.1±δ
478.99
0.01380
1.19
0.78±0.12
Page 143
Chapter VI. Thermal Conductivity of PLS Compounds
Sr3.6La0.4Nb4O14.2±δ
484.91
0.01390
1.18
0.95±0.09
Sr3.4La0.6Nb4O14.3±δ
490.84
0.01390
1.17
0.96±0.11
Sr3.2La0.8Nb4O14.4±δ
496.77
0.01400
1.16
0.98±0.07
It is evident from Table 6.5 that after substitution, the atomic mass decreased
in the case of La4-xSrxTi4O14La4Ti4-xNbxO14+
±δ
±δ
and increased in the case of La4Ti4-xTaxO14+
and Sr4-xLaxNb4O14+
±δ
±δ,
compared to the pure La4Ti4O14 and
Sr4Nb4O14. Assuming E to be nearly constant for all the compositions [11, 12], the
thermal conductivity will increase in the case of La4-xSrxTi4O14the case of La4Ti4-xTaxO14+
±δ,
La4Ti4-xNbxO14+
±δ
±δ
and decrease in
and Sr4-xLaxNb4O14+
±δ
due to
increasing molecular weight. After the substitution of Sr in La4Ti4O14, the thermal
conductivity became less temperature dependent and increased slightly with
increasing temperature. In the case of Ta, Nb and La substituted ceramics, the
thermal conductivity decreased with the increasing substitution following the same
temperature independency. After reduction, the thermal conductivity decreased
further.
In the case of substituted La4Ti4O14 and Sr4Nb4O14, a high density of
nanoscale intergrowths and secondary phase were found inside the microstructure.
So the behaviour of thermal conductivity depends upon the atomic mass difference
and the nanoscale intergrowths/secondary phase. Also, the experimental thermal
conductivity values are lower than the minimum thermal conductivity value. This
decrease in thermal conductivity by substituting different elements can be explained
by the fact that substituted elements incorporates nanoscale intergrowths in the
structure to accommodate non stoichiometry. The phonons are scattered by these
nanoscale intergrowths to reduce thermal conductivity compared to the pure
Page 144
Chapter VI. Thermal Conductivity of PLS Compounds
compound (and theoretical minimum thermal conductivity values). The combined
effect of phonon scattering from the nanoscale intergrowths and the substituted
atoms resulted in a change in thermal conductivity.
In order to estimate the effect of nanoscale intergrowths/secondary phase on
thermal conductivity, the difference between the calculated thermal conductivity
(given in Table 6.5) and the measured thermal conductivity was calculated. This
difference was found to be negligible in the case of pure compounds. As the amount
of substituted elements increased, the difference increased to ~ 20% which was the
result of nanoscale intergrowths/secondary phase in lowering thermal conductivity.
The Debye temperature (θD) for La4Ti4O14 and Sr4Nb4O14 based
compositions was calculated by the following relationship [13] and the data is shown
in the Table 6.6.
θD =
V
(6.3)
Where h is the Plank’s constant and V is the velocity of sound in solids. The
phonon mean free path (η) is defined as the average distance a phonon travels before
being scattered. η can be calculated by using the high temperature limit of Debye
equation and the Dulong Petit equation [2]. A temperature of 973 K was chosen for
the calculation.
η
(6.4)
where κ is the real time thermal conductivity value. η for all the air annealed
ceramics is presented in the Table 6.6.
Page 145
Chapter VI. Thermal Conductivity of PLS Compounds
Table 6.6: Phonon mean free path and Debye temperature for air annealed La4Ti4O14
and Sr4Nb4O14 based compositions
Composition
Phonon mean free path η Debye Temperature (K)
(Å)
La4Ti4O14±δ
2.7
568
La3.8Sr0.2Ti4O13.9±δ
2.7
572
La3.6Sr0.4Ti4O13.8±δ
2.8
573
La3.4Sr0.6Ti4O13.7±δ
2.9
578
La3.2Sr0.8Ti4O13.6±δ
2.9
581
La4Ti3.9Ta0.1O14.05±δ
2.3
565
La4Ti3.8Ta0.2O14.1±δ
2.2
561
La4Ti3.6Ta0.4O14.2±δ
2.2
550
La4Ti3.4Ta0.6O14.3±δ
2.4
542
La4Ti3.8Nb0.2O14.1±δ
2.7
567
La4Ti3.6Nb0.4O14.2±δ
2.2
564
La4Ti3.4Nb0.6O14.3±δ
2.3
560
Sr4Nb4O14±δ
2.9
604
Sr3.8La0.2Nb4O14.1±δ
2.3
606
Sr3.6La0.4Nb4O14.2±δ
2.3
605
Sr3.4La0.6Nb4O14.3±δ
2.2
605
Sr3.2La0.8Nb4O14.4±δ
2.0
603
The η for air annealed La4Ti4O14 based compositions increased slightly from
~ 2.7 Å to ~ 3.0 Å after Sr substitution while it decreased to ~ 2.3 Å after Ta/Nb
Page 146
Chapter VI. Thermal Conductivity of PLS Compounds
substitution. For air annealed Sr4Nb4O14, η was estimated to be ~ 3.0 Å which
decreased to ~ 2.0 Å after La substitution.
In perovskites (e.g. SrTiO3) the lattice parameter is 3.95 Å (XRD PDF # 00035-0734) while in PLS structure, the layers are presented along (110) cut of the unit
cell. From this information, the thickness of individual layer for La4Ti4O14 was
calculated to be ~ 2.7 Å while the separation between the perovskite blocks in the
unit cell was found to be ~ 2.04 Å as shown in Figure 6.19. The η values (Table 6.6)
are similar to the thickness of individual perovskite layer (2.7 Å) and the separation
between the perovskite blocks (2.04 Å). From this information, it can be concluded
that the phonon mean free path approaches the atomic dimensions which explains the
temperature independency of thermal conductivity for PLS materials.
Figure 6.19: Schematic representation of a typical non distorted La4Ti4O14 projected along
a axis
Page 147
Chapter VI. Thermal Conductivity of PLS Compounds
6.6
CONCLUSION
Acceptor and donor substituted La4Ti4O14 and donor substituted Sr4Nb4O14
were synthesised by spark plasma sintering. The effect of oxidation/reduction and
compositional non stoichiometry were studied on thermal properties. Sr/Ta
substitution and reduction produced nanoscale intergrowths to accommodate the
compositional non-stoichiometry. An extremely low thermal conductivity value of
~ 0.93 W/m.K (at 573 K) was observed in the case of reduced La4Ti3.4Ta0.6O14.3±δ
ceramic as a result of the nanoscale intergrowths. This value is amongst the lowest
value ever reported for the untextured oxide materials. Nb substitution in La4Ti4O14
and La substitution in Sr4Nb4O14 did not produce nanoscale intergrowths but
produced LaNbO4 as a separate phase.
Page 148
Chapter VI. Thermal Conductivity of PLS Compounds
REFERENCES
[1] A. Sakai, T. Kanno, K. Takahashi, Y. Yamada, H. Adachi, Journal of Applied
Physics, 108 (2010) 103706.
[2] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
[3] A. Sakai, T. Kanno, K. Takahashi, A. Omote, H. Adachi, Y. Yamada, X.D. Zhou,
Journal of the American Ceramic Society, 95 (2012) 1750-1755.
[4] B. Liu, J.Y. Wang, F.Z. Li, Y.C. Zhou, Acta Materialia, 58 (2010) 4369-4377.
[5] Z.G. Liu, J. H. Ouyang, Y. Zhou, Bulletin of Materials Science, 32 (2009) 603606.
[6] Y. Shen, D.R. Clarke, P.A. Fuierer, Applied Physics Letters, 93 (2008) 102907.
[7] B.S. Mitchell, Transport Properties of Materials, in: An Introduction to Materials
Engineering and Science, John Wiley & Sons, inc. , Canada, 2004, pp. 320.
[8] G. Li, J.Y. Yang, Y. Xiao, L.W. Fu, J.Y. Peng, Y. Deng, P.W. Zhu, H.X. Yan, J.
Electron.Mater. 42 (2013) 675–678.
[9] M.R. Winter, D.R. Clarke, Journal of the American Ceramic Society, 90 (2007)
533-540.
[10] C.G. Levi, Curr Opin Solid St M, 8 (2004) 77-91.
[11] D.R. Clarke, Surf Coat Tech, 163 (2003) 67-74.
[12] R. Korobko, S.K. Kim, S. Kim, S.R. Cohen, E. Wachtel, I. Lubomirsky,
Advanced Functional Materials, 23 (2013) 6076-6081.
[13] R. Freer, Journal of Materials Science, 16 (1981) 3225-3227.
Page 149
Chapter VII. Thermoelectric Properties of co-doped CoSb3
Chapter VII. Thermoelectric Properties of codoped CoSb3
7.1
INTRODUCTION
Skutterudites such as cobalt antimonide (CoSb3) [3] have been the focus of
the research for intermediate temperature thermoelectric applications (25 oC to 550
o
C) [1, 2]. CoSb3 has a melting point of 873 oC [4] with high carrier concentration
and Seebeck coefficient (~ 60µV/K) [5]. However, due to strong covalent bonding,
the intrinsic thermal conductivity is ~ 10 W/m.K which is a high value for
thermoelectric applications [6]. CoSb3 has cubic structure (Space group Im3) with
two naturally formed vacant atomic cages per unit cell. These cages favour the
PGEC approach and can be partially filled with a variety of atoms like rare earths or
alkali metals [7]. These atoms are loosely bonded and rattle inside the cage which
causes phonons to scatter and effectively decreases the thermal conductivity without
degrading the electrical properties [8, 9]. Substitution of Te for Sb is an effective way
of improving the electrical properties of CoSb3. A significant reduction in electrical
resistivity for Co4Sb12-xTex was observed as compared to pure CoSb3 because Te
serves as an electron donor in the CoSb3 structure [10].
Ytterbium (Yb) is an effective dopant in the CoSb3 structure due to its high
atomic mass and (173.04 amu) and small ionic size (0.0858nm). By partially filling
the cages in CoSb3 structure with Yb, the thermal conductivity is significantly
decreased (~ 4 W/m.K) [11] while slightly affecting the Seebeck coefficient and
electrical resistivity [6].
Page 150
Chapter VII. Thermoelectric Properties of co-doped CoSb3
In the present work, we used the concept of PGEC for CoSb3 to reduce
thermal conductivity while maintaining the electrical properties. Previous work on
mixed stuffing and substitution reported lattice thermal conductivity of 1.5 W/m.K
[12] and 1.3 W/m.K [13]. These reports utilized indium (In) stuffing in germanium
(Ge) substituted and nickel (Ni) substituted CoSb3 materials respectively. For the
first time ytterbium stuffed and tellurium substituted CoSb3 was synthesized in order
to decrease thermal conductivity without affecting electrical conductivity and
Seebeck coefficient.
7.2
EXPERIMENTAL DETAILS
Pure and Yb and Te doped CoSb3 based compositions were synthesised by
mechanical alloying. Commercially available powders of Co (99.8% pure), Sb
(99.5% pure), Te (99.999% pure) and Yb (99.8% pure) were mixed in the
stoichiometric
ratios
to
obtain
CoSb3
(CS),
CoSb2.85Te0.15
(CST)
and
Yb0.075CoSb2.85Te0.15 (YCST). Mixing of powders was carried out inside an Ar filled
glove box in steel jars to avoid oxidation of the powders. The steel jars were loaded
into the planetary ball mill and rotated at a speed of 350 rpm for 40 hours [14]. The
mechanically alloyed powders were sintered into 20 millimetre discs by SPS at 600
o
C for 5 minutes in an argon environment. A heating rate of 100oC/min was used. All
the samples were more than 98% dense as measured by the Archimedes principle.
The samples were cut into 3 x 3 x 15 mm bars for electrical resistivity and Seebeck
coefficient measurement via temperature differential and four point probe methods in
lab-made apparatus under vacuum.
Page 151
Chapter VII. Thermoelectric Properties of co-doped CoSb3
7.3
RESULTS AND DISCUSSIONS
Figure 7.1 shows XRD of sintered CoSb3 based materials. The diffraction
peaks indicate that the major phase is CoSb3 (XRD PDF card: 76-0470) with slight
amounts of impurity phases which are CoSb2 (XRD PDF card: 65-4102) and CoTe2
(XRD PDF card: 89-2091). These results are consistent with the previous studies,
[10] which indicate that the formation of a secondary phase is a common problem.
All of the peaks of the major phase in the Te substituted material are shifted towards
low angle, which indicates that the lattice parameter has increased. This indicates
that the Te and Yb atoms were accommodated into the structure of CoSb3.
Figure 7.1: (a) X ray diffraction patterns for CoSb3 based materials; (b) SEM micrograph of CoSb3
SEM images showed that the CoSb3 had high density (~ 98%) and consisted
of angular grains ~ 0.8 µm. For the materials doped and substituted with Yb and Te,
Page 152
Chapter VII. Thermoelectric Properties of co-doped CoSb3
the grains size decreased to 0.27 µm and 0.19 µm for CST and YCST respectively
without affecting the grain shape.
Figure 7.2(a) shows electrical resistivity data for the CoSb3 based materials.
The electrical resistivity of CoSb3 decreased with increasing temperature consistent
with its semiconducting properties [15]. A considerable decrease in electrical
resistivity was observed upon Te addition compared to that of pure CS. The
electrical resistivity increased with increasing temperature indicating metallic like
behaviour [16]. As mentioned earlier, Te atoms serve as electron donors in the CS
lattice, thus providing more electrons to the structure leading to decreased electrical
resistivity. This increased carrier concentration raises the Fermi level giving rise to a
degenerate semiconductor [17]. Yb addition to CST further reduced electrical
resistivity slightly as shown in Figure 7.2(a). This indicates that Yb donates its
outermost electrons, which further decreases electrical resistivity [18].
Page 153
Chapter VII. Thermoelectric Properties of co-doped CoSb3
Figure 7.2: Variation of: (a) electrical resistivity; (b) Seebeck coefficient with temperature for
CoSb3 compositions
Figure 7.2(b) shows Seebeck coefficient data for the CoSb3 based materials.
At room temperature CS behaves as an n type semiconductor but at temperatures
higher than 600 K, it behaves as a p type semiconductor, which is consistent with
other reports [19]. This is because the mobility of the holes produced by the strong
intrinsic conduction is much higher in the temperature range of more than 600 K
Page 154
Chapter VII. Thermoelectric Properties of co-doped CoSb3
than the electron mobility produced by the weak extrinsic conduction [20]. Addition
of Yb and Te at the same time to CoSb3 resulted in a higher negative Seebeck value
that increased with increasing temperature. A maximum value of ~ 160 µV/K was
achieved at 600 K. This increase is gradual and the value of Seebeck coefficient
remains nearly constant in the temperature range of ~ 600-800 K. Considering the
Fermi Dirac statistics, the Seebeck coefficient values can be expressed as [21]
S
kB
e
 F1  

 2
  
 F0  

(2.6)
Where kB is the Boltzmann constant, e is the unit charge, Fn(  ) is the nth order
Fermi integral and  is the reduced Fermi energy. Since the Fermi energy decreases
with temperature, the (negative) Seebeck coefficient increases with increasing
temperature. In the temperature range of 600-800 K, the value of the Seebeck
coefficient remains nearly constant which is attributed to the mixed conduction [22].
The combined effect of both the charge carriers in CST and YCST results in the
Seebeck coefficient being nearly constant.
The most dramatic impact of Yb stuffing and Te substitution in CoSb3 is on
its thermal conductivity. In Figure 7.3(a) the total thermal conductivity (κT) of YCST
is plotted against temperature. For comparison, κT of CS and CST are also plotted. In
the case of pure CoSb3 the κT decreases with increase in temperature and reaches a
minimum at ~ 550 K. After reaching the minimum the value of κT increases with
increase in temperature. This increase of thermal conductivity at higher temperature
is related to the bipolar effect which will be explained later.
Page 155
Chapter VII. Thermoelectric Properties of co-doped CoSb3
From the data it is obvious that the simultaneous addition of Yb and Te is
more effective than substituting Te alone into CoSb3. A minimum value of 2.19
W/mK was observed at ~ 550 K for YCST. This value is lower than the one reported
for single crystal CoSb3 (4 W/m.K) [23], and Yb0.26Co4Sb12/GaSb nanocomposite
(2.6W/m.K) [22]. Our value is comparable to Yb doped nanostructured CoSb3
(2.15 W/m.K) [24], which indicates that instead of employing costly nanostructuring
processing, a very low value of thermal conductivity can be achieved by
simultaneous stuffing and substitution.
Figure 7.3: Variation of: (a) total thermal conductivity; (b) lattice thermal conductivity with
temperature for CoSb3 based compositions
Page 156
Chapter VII. Thermoelectric Properties of co-doped CoSb3
In order to study the effect of Yb stuffing and Te substitution on lattice
thermal conductivity (κL), the lattice thermal conductivity was calculated by
subtracting the electronic contributions to the thermal conductivity (κe) from total
thermal conductivity. κe was calculated by using Weidman- Franz Law which is
written as
κelec= LTσ
(2.4)
Where L is the Lorenz number, T is the absolute temperature and σ is the electrical
conductivity. Assuming CoSb3 is a homogeneous material with a parabolic band
dominated by acoustic phonon scattering at low temperatures (κLattice α T-1), the
Lorenz number is given as [21]
2
2
 k   3F  F2    4 F1   
L   B   0
2

F0  
 e  

(2.5)
 increases by the addition of Yb in CoSb2.85Te1.5 which indicates that the CoSb3 is
still a degenerate semiconductor which is in consistent with the electrical resistivity.
The calculated L values for CS, CST and YCST were 2.3x10-08 V2/K2, 1.8x10-08
V2/K2 and 1.9x10-08 V2/K2 respectively which are in good agreement with literature
[17].
Figure 7.3(b) shows the variation of the lattice thermal conductivity with
temperature. An almost 70% reduction in lattice thermal conductivity of CS was
produced by the addition of Yb and Te to CoSb3, reaching a minimum value of 1.17
W/m.K. This is lower than the one reported by Malik et. al (1.5 W/m.K) for In
Page 157
Chapter VII. Thermoelectric Properties of co-doped CoSb3
stuffed and Ge doped CoSb3 [11] and (~ 1.3W/m.K) for In stuffed and Ni doped
CoSb3 [12]. This value is even lower than double stuffed CoSb3 with Yb and Ba (2.2
W/m.K) [8]. However, the lattice thermal conductivity is still higher than the
calculated minimum thermal conductivity of 0.66 W/m.K for YCST calculated using
[25]
(6.1)
Where E is the elastic modulus taken as 148 GPa [26].
In order to investigate the effect of Yb and Te on CoSb3, the phonon mean
free path (η) of all the 3 materials were calculated by using [27]
η
(6.4)
Where κ is the lattice thermal conductivity and V is the velocity of sound in the
solids calculated through elastic modulus and density data. Substituting the
respective values of κ, Ω, kB and V into the equation, η for CS, CST and YCST were
estimated to be ~ 1.57 nm, 0.98 nm and 0.76 nm respectively. From these values it is
evident that the phonon mean free path was almost halved by the addition of Yb and
Te (compared to CS) which corresponds to the separation between the rare earth
elements rattling inside a crystal (~ 8 Å) [28].
Page 158
Chapter VII. Thermoelectric Properties of co-doped CoSb3
Figure 7.4: Components of total thermal conductivity with T-1 for YCST
The minor charge carriers which start operating in the higher temperature
region (550 K to 750 K) not only affect the Seebeck coefficient, but also increase
thermal conductivity. Electrons - holes pairs are generated by the absorption of
energy at the hot end. These pairs recombine by releasing energy when they move to
the cold end and thus increase thermal conductivity. This phenomenon is called as
bipolar diffusion. So, the total thermal conductivity can now be presented by
κTotal = κelectronic + κLattice + κBipolar
(7.1)
In order to determine the contribution of bipolar diffusion to thermal
conductivity in the intrinsic conduction range, a method proposed by Kitgawa et al.
is employed [29]. Figure 7.4 shows variation of κLattice (κLattice = κElectronic - κTotal) and
κTotal for YCST as a function of T-1. κTotal can be separated into κelectronic, κLattice and
Page 159
Chapter VII. Thermoelectric Properties of co-doped CoSb3
κBipolar. Since at low temperatures, the lattice thermal conductivity is proportional to
T-1 according to the assumption made, the κLattice was estimated by extrapolating the
linearity to high temperatures. As can be seen, the κLattice is the dominant contribution
to total thermal conductivity at low temperatures but at higher temperatures, a
prominent increase in κBipolar was observed. From the above discussion it can be
concluded that above the bipolar transition temperature, intrinsic conduction
dominates.
Figure 7.5 shows variation of zT with temperature calculated from measured
Seebeck coefficient, electrical resistivity and total thermal conductivity. zT value
increases with increase in temperature for doped compositions and reaches a value of
~ 0.70 at 600 K for YCST. This increase in zT value is due to the low value of lattice
thermal conductivity produced by the partial filling by Yb atoms of the cage like
structure of CoSb3.
Figure 7.5: Temperature dependence of zT for CoSb3 based compositions
Page 160
Chapter VII. Thermoelectric Properties of co-doped CoSb3
7.4
CONCLUSION
Combined Yb stuffed and Te substituted CoSb3 was synthesized by
mechanical alloying and spark plasma sintering. Lattice thermal conductivity was
significantly reduced to a very low value of 1.17 W/m.K by the addition of Yb atoms
into CoSb3-xTex without significantly affecting Seebeck coefficient and electrical
resistivity. This value is comparable to those of produced by the costly processing of
nanostructured materials. A zT value of ~ 0.70 was obtained at 600 K.
Page 161
Chapter VII. Thermoelectric Properties of co-doped CoSb3
REFERENCES
[1] G. Li, J.Y. Yang, Y. Xiao, L.W. Fu, J.Y. Peng, Y. Deng, P.W. Zhu, H.X. Yan,
Journal of Electronic Materials, 42 (2013) 675-678.
[2] J.J. Zhang, B. Xu, L.M. Wang, D.L. Yu, Z.Y. Liu, J.L. He, Y.J. Tianb, Applied
Physics Letters, 98 (2011) 072109.
[3] A. Harnwunggmoung, K. Kurosaki, H. Muta, S. Yamanaka, Applied Physics
Letters, 96 (2010) 202107.
[4] H.Y. Geng, S. Ochi, J.Q. Guo, Applied Physics Letters, 91 (2007) 022106.
[5] W.-S. Liu, B.-P. Zhang, J.-F. Li, L.-D. Zhao, Journal of Physics D: Applied
Physics, 40 (2007) 566-572.
[6] G.S. Nolas, M. Kaeser, R.T. Littleton, T.M. Tritt, Applied Physics Letters, 77
(2000) 1855.
[7] B.C. Sales, D. Mandrus, R.K. Williams, Science, 272 (1996) 1325-1328.
[8] X. Shi, H. Kong, C.P. Li, C. Uher, J. Yang, J.R. Salvador, H. Wang, L. Chen, W.
Zhang, Applied Physics Letters, 92 (2008) 182101.
[9] H. Li, X.F. Tang, Q.J. Zhang, C. Uher, Applied Physics Letters, 93 (2008)
252109.
[10] W.S. Liu, B.P. Zhang, L.D. Zhao, J.F. Li, Chemistry of Materials, 20 (2008)
7526-7531.
[11] G. S. Nolas M. Kaeser, R. T. Littleton, T. M. Tritt, Applied Physics Letters, 77
(2000) 1855.
[12] R. Chandra Mallik, E. Mueller, I.-H. Kim, Journal of Applied Physics, 111
(2012) 023708.
Page 162
Chapter VII. Thermoelectric Properties of co-doped CoSb3
[13] J. Jae-Yong, P. Kwan-Ho, K.I.M. Il-Ho, S.-M. Choi, W.-S. Seo, Journal of the
Korean Physical Society, 57 (2010) 773.
[14] J. Yang, Y. Chen, J. Peng, X. Song, W. Zhu, J. Su, R. Chen, Journal of Alloys
and Compounds, 375 (2004) 229-232.
[15] J.X. Zhang, Q.M. Lu, K.G. Liu, L. Zhang, M.L. Zhou, Materials Letters, 58
(2004) 1981-1984.
[16] Y. Zhu, H. Shen, L. Zuo, H. Guan, Solid State Communications, 151 (2011)
1388-1393.
[17] W.-S. Liu, B.-P. Zhang, J.-F. Li, H.-L. Zhang, L.-D. Zhao, Journal of Applied
Physics, 102 (2007) 103717.
[18] A. Zhou, L. Liu, P. Zhai, W. Zhao, Q. Zhang, Journal of Applied Physics, 109
(2011) 113723.
[19] W.-S. Liu, B.-P. Zhang, J.-F. Li, L.-D. Zhao, Journal of Physics D: Applied
Physics, 40 (2007) 6784-6790.
[20] M.J. Kim, I.H. Kim, Met Mater Int, 16 (2010) 459-463.
[21] H. Anno, K. Matsubara, Y. Notohara, T. Sakakibara, H. Tashiro, Journal of
Applied Physics, 86 (1999) 3780-3786.
[22] Z. Xiong, X.H. Chen, X.Y. Huang, S.Q. Bai, L.D. Chen, Acta Materialia, 58
(2010) 3995-4002.
[23] T. Caillat, A. Borshchevsky, J.P. Fleurial, Journal of Applied Physics, 80 (1996)
4442-4449.
[24] H. Li, X. Tang, X. Su, Q. Zhang, C. Uher, Journal of Physics D: Applied
Physics, 42 (2009) 145409.
[25] C.G. Levi, Curr Opin Solid St M, 8 (2004) 77-91.
Page 163
Chapter VII. Thermoelectric Properties of co-doped CoSb3
[26] X.Q. Yang, P.C. Zhai, L.S. Liu, Q.J. Zhang, Journal of Applied Physics, 109
(2011) 123517.
[27] T.D. Sparks, P.A. Fuierer, D.R. Clarke, Journal of the American Ceramic
Society, 93 (2010) 1136-1141.
[28] B.C. Sales, D. Mandrus, B.C. Chakoumakos, V. Keppens, J.R. Thompson,
Physical Review B, 56 (1997) 15081-15089.
[29] H. Kitagawa, M. Wakatsuki, H. Nagaoka, H. Noguchi, Y. Isoda, K. Hasezaki,
Y. Noda, Journal of Physics and Chemistry of Solids, 66 (2005) 1635-1639.
Page 164
Chapter VIII. Conclusions and Future Work
Chapter VIII. Conclusions and Future Work
8.1
CONCLUSION
In this research, the effect of substitution and oxidation-reduction on the
microstructure and thermoelectric properties was studied. Research was conducted
on oxide and non-oxide materials. For oxide ceramics, La4Ti4O14 and Sr4Nb4O14
were studied. La4-xSrxTi4O14-
±δ
(x=0, 0.2, 0.4, 0.6 and 0.8), La4Ti4-xTaxO14+
0.1, 0.2, 0.4 and 0.6), La4Ti4-xNbxO14+
Sr4-xLaxNb4O14+
±δ
±δ
±δ
(x=0,
(x=0, 0.2, 0.4 and 0.6) and
(x=0, 0.2, 0.4, 0.6 and 0.8) were prepared by solid state reaction.
The grain shape and size decreased as compared to the pure compositions.
Acceptor substituted La4Ti4O14 produced nanoscale intergrowths of 5 layer
PLS ceramic and the amount increased after reduction while donor substitution
produced intergrowths of 3 layer PLS ceramic which increased after oxidation and
decreased after reduction. This was observed by XRD when additional peaks
appeared. La4Ti4-xNbxO14+
±δ
and Sr4-xLaxNb4O14+
±δ
did not produce intergrowths
but a second phase LaNbO4.
Microstructural and thermal properties were characterised for these materials.
Reduction produced some planar defects arising in the pure compositions. TEM
confirmed the existence of intergrowths of 5 layer region in La4-xSrxTi4O14-
±δ
existence of 3 layer regions could not be determined in La4Ti4-xTaxO14+
This was
±δ.
but the
due to the complex nature of 3 layer PLS ceramics. Formation of LaNbO4 as a
separate phase was also confirmed in La4Ti4-xNbxO14+
±δ and
Sr4-xLaxNb4O14+
±δ.
The
Page 165
Chapter VIII. Conclusions and Future Work
solubility limit of La in polycrystalline Sr4-xLaxNb4O14+
La4Ti4-xNbxO4+
±δ
±δ
and Nb in polycrystalline
is x=0.2.
The thermal properties of the PLS compounds were characterized. All PLS
materials showed anomalously low thermal conductivity values. Heavy element
substitution (La, Nb and Ta) reduced thermal conductivity while light element (Sr)
increased thermal conductivity with increasing amount of substitution due to the
mass contrast. Apart from the mass contrast, the thermal conductivity decreased with
increasing amount of nanoscale intergrowths which were generated to accommodate
compositional non-stoichiometry. A ~ 20% decrease in the theoretical minimum
thermal conductivity was achieved due to the nanoscale intergrowths. A very low
thermal conductivity value of ~ 0.93 W/m.K was observed in Ta substituted
La4Ti4O14.
For non-oxide ceramics, CoSb3 was chosen due to its cage-like structure.
CoSb3 stuffed with Yb and substituted with Te (YbyCoSb3-xTex) was synthesized by
mechanical alloying and spark plasma sintering. Electrical and thermal properties
were characterized for pure and doped material. A Seebeck coefficient value of
~ 160 µV/K was obtained at ~ 600-800 K for Yb0.075CoSb2.85Te0.15. Electric
resistivity dropped from ~ 1000 µΩm for pure CoSb3 to ~ 9 µΩm for
Yb0.075CoSb2.85Te0.15. Lattice thermal conductivity was significantly reduced to a
very low value of 1.17 W/m.K by the addition of Yb atoms into CoSb2.85Te0.15
without significantly affecting Seebeck coefficient and electrical resistivity. This
value is comparable to those of produced by the costly processing of nanostructured
materials. A zT value of ~ 0.70 was obtained at 600K.
Page 166
Chapter VIII. Conclusions and Future Work
This research has shown that by engineering the defect chemistry and
compositional non-stoichiometry of thermoelectric materials, it is possible to
significantly reduce their thermal conductivity without compromising their electrical
properties.
Page 167
Chapter VIII. Conclusions and Future Work
8.2
FUTURE WORK
8.2.1
Grain Size Effect
Reducing grain size can significantly reduce thermal conductivity. This effect
on thermal conductivity is not studied in PLS compounds. Hence, it would be very
useful to compare the effect of grain size on thermal properties of these A4B4O14
ceramics. This effect can also be related to investigate the electric properties. Spark
Plasma Sintering helps to minimise the grain size of the nano-sized ceramics during
sintering.
8.2.2 Textured PLS thermoelectrics
As the PLS compounds have plate-like grains and with the help of SPS, these
compounds can be textured. A two-step texturing method has been reported in
literature to synthesise dense ceramics. Exploring the PLS materials for texturing
with different substitution and studying the effect of different variables to control the
texturing and grain growth. This can lead to a systematic study to improve
thermoelectric properties in PLS materials.
8.2.3 Different PLS compounds
In this research 4 layer PLS compounds have studied and it has been
shown that the disorder helps in reduction of thermal conductivity. Other members
Page 168
Chapter VIII. Conclusions and Future Work
of the PLS family i.e. 2 and 3 layer PLS compound (LaNbO4, La3Ti2TaO11, LaTaO4
etc.) have similar layered crystal structure. There are no known reports of thermal
conductivity for these compounds. It may be possible to obtain low thermal
conductivity values in these materials and explore the effect of heavy element
substitution on A and/or B site.
8.2.4 Melt Spinning of Skutterudite
Melt spinning of skutterudite has been reported to yield good thermoelectric
properties. The composition Yb0.075CoSb2.85Te0.15 has not been explored by melt
spinning. Since this composition has yielded a zT value of ~ 0.7, synthesising this
composition with melt spinning can lead to improve zT value of this material. Since
Yb0.075CoSb2.85Te0.15 is an n type compound, its p type counterpart can also be
studied by melt spinning and SPS.
Page 169
List of my publications
List of my publications
1. H. Porwal, P. Tatarko, S. Grasso, J. Khaliq, I. Dlouhý and M. J. Reece, ‘‘Graphene
reinforced alumina nano-composites’’, Carbon, 2013, 64, 359-369
2. S. Grasso, N. Tsujii, Q. H. Jiang, J. Khaliq, S. Maruyama, M. Miranda, K. Simpson,
T. Mori and M. J. Reece, ‘‘Ultra low thermal conductivity of disordered layered ptype bismuth telluride’’, Journal of Materials Chemistry C, 2013, 1,2362-2367
3. G.Viola, H. Ning, X. Wei, M. Deluca, A. Adomkevicius, J. Khaliq, M. J. Reece and
H. Yan, ‘‘Dielectric relaxation, lattice dynamics and polarization mechanisms in
Bi0.5Na0.5TiO3-based lead-free ceramics’’, Journal of Applied Physics, 2013, 114,
014107
4. J. Khaliq, Q Jiang, J. Yang, K. Simpson, H. Yan and M. J. Reece, ‘‘Utilizing the
phonon glass electron crystal concept to improve the thermoelectric properties of
combined Yb-stuffed and Te-substituted CoSb3’’, Scripta Materialia, 2014, 72-73,
63-66
5. Q. Jiang, H. Yan, J. Khaliq, H. Ning, S. Grasso, and M. J. Reece, ‘‘Large ZT
enhancement in hot forged nanostructured p-type Bi0.5Sb1.5Te3 bulk alloys, Journal of
Materials Chemistry A, 2014, 2, 5785
6. J. Khaliq, C. Li, K. Chen, B. Shi, H. Ye, K. Simpson, H. Yan, A.M. Grande and M.
J. Reece, Reduced thermal conductivity induced by nanoscale intergrowths in
perovskite like layered structure La2Ti2O7, (Under Preparation)
Page 170